201901 Hokkaido Conference Proceedings (SS)

ISSSM-0388 Balancing a Bike-Sharing System Through Dynamic Pricing Strategy Peng-Sheng Youa, Ta-Cheng Chenb, Yi-Chih Hsiehc a Department of Business Administration, National Chiayi University, Chiayi, Taiwan b Department of M-Commerce and Multimedia Applications, Asia University, Taichung, Taiwan c Department of Industrial Management, National Formosa University, Yunlin, Taiwan E-mail: [email protected] a, [email protected] b, [email protected] 1. Background Bicycle-sharing systems are commonly established at dispersed locations to create their rental service networks. these systems commonly allow customers to pick up bicycles from one station and return them to a different one. However, as time goes by, this service may face the following two problems: (1) For stations with bike outflows greater than their inflows, they may run the risk of lost sales due to insufficient bikes; (2) For stations with bike inflows greater than their outflows, they may run the risk of idle bikes and insufficient bike slots (Sayarshad et al., 2012). Neither one is preferred by the rental companies since either one will reduce service quality or revenues. Bicycle firms can improve the imbalance problem using the operator-based or the user-based rebalancing approaches to improve service quality or profits. A number of mathematical models have been proposed to deal with the bicycle imbalance problems (e.g., Benchimol et al., 2011; Chemla et al., 2013; Ho and Szeto, 2014; Song and Carter, 2008; You et al., 2017; Raviv and Kolka, 2013). The mentioned literature dealt with the imbalance problem in terms of supply management approaches (e.g., Song and Carter, 2008, Erdoğan et al., 2015). From an economic point of view, in addition to the supply management methods, pricing approaches can also be applied to balance the gap between demand and supply. However, to the best of our knowledge, few researchers have attempted to overcome the problem using the demand management approaches. Thus, this paper attempted to use dynamic pricing approach to deal with this problem. 2. Methods This paper developed a mathematical programming model with the operator-based and the user-based rebalancing approaches to deal with the imbalance problem. The purpose of the proposed model is to maximize the total profit, total rental revenues minus total penalty costs from unmet demand and empty bike transport, over a planning horizon by determining the dynamic rental price and deployments. The proposed approach will be applied to the analysis of public bike systems in Taiwan. Sensitivity analyses were also be conducted to investigate the effects of various system parameters on decisions and service qualities. 448

3. Results This paper established a constrained mathematical model to deal with a rental pricing problem for bicycle rental companies. We develop a heuristic algorithm based on the linear programming and an evolutionary algorithm to efficiently solve the problem. Computational results shows that the proposed approach can find optimal solutions for small-scale problems and can produce satisfactory solutions within a reasonable CPU time. Sensitivity analyses were also conducted to investigate the effects of various system parameters. Keywords: Bike sharing system, rebalancing, dynamic pricing, mathematical programming 4. References Benchimol, M., Benchimol, P., Chappert, B., Taille, A.D.L., Laroche, F., Meunier, F. & Robinet, L. (2011) Balancing the stations of a self-service bike hire system, RAIRO Operations Research, 45, 37–61. Chemla, D., Meunier, F. & Wolfler Calvo, R. (2013) Bike sharing systems: solving the static rebalancing problem, Discrete Optimization 10, 120-146. Erdoğan, G, Battarra, M. & Wolfler Calvo, R. (2015) An exact algorithm for the static rebalancing problem arising in bicycle sharing systems, European Journal of Operational Research 245, 667-79. Ho, S.C. & Szeto, W.Y. (2014) Solving a static repositioning problem in bike-sharing systems using iterated tabu search, Transportation Research E 69 (2014), 180–198. Sayarshad, H.R., Tavassoli, S., & Zhao, F. (2012) A multi-periodic optimization formulation for bike planning and bike utilization, Applied Mathematical Modelling, 36, 4944–4951. Song, D.P. & Carter, J. (2008) Optimal empty vehicle redistribution for hub-and-spoke transportation systems, Naval Research Logistics, 55, 156-71. You, P.S., Lee, P.J. & Hsieh, Y.C. (2017) An artificial intelligent approach to the bicycle repositioning problems, Engineering Computations, 34, 145-163. Raviv, T. & Kolka, O. (2013) Optimal inventory management of a bike-sharing station, IIE Transactions 45, 1077-1093. 449

ISSSM-0395 Antecedents and Consequences of the Competitive Advantage for SMEs in Thailand Araya Uengpaiboonkit Management Department, Faculty of Management Technology, Rajamangala University of Technology Isan, Surin Campus, Thailand E-mail: [email protected] Abstract This study aims to (1) investigate antecedent the factors influencing the competitive advantage and (2) explore influence of competitive advantage upon performance. The quantitative research study was conducted using a questionnaire survey, the researcher defined 128,516 the representative population for this research as a small and medium enterprises in Burirum, Surin and Srisaket province in Thailand. Totaling 1,200 subjects which were selected by multi-stage sampling method. Data was analyzed by Structural Equation Modeling. Findings from the research suggested that: (1) Intellectual Capital, Innovation, and Dynamic Capabilities had direct and positive influence on the competitive advantage. and (2) Competitive Advantage showed direct and positive influence on performance. In addition, effective performance was indirectly influenced by Intellectual Capital, Innovation, and Dynamic Capabilities with the research hypotheses. Keywords: Competitive Advantage, Small and Medium Enterprises 1. Background Small and Medium Sized Enterprises: SMEs are currently receiving considerable interest from related parties, especially South Korean and multinational enterprises 2016 totaling 2,844,757 by small and medium enterprises (SMEs). Of 2,736,744, accounting for 99.53 percent, growing by 0.76 percent, bringing the total gross domestic product (GDP) to 5,212,004 million baht or 39.6 percent of GDP (Office of Small and Medium Enterprises Promotion, 2017). SMEs are a way to generate revenue and add value to the government as a whole (Boone and Kurtz, 2010). According to the SMEs employment report for 2015, employment in small and medium enterprises is estimated at 10,501,166 or 80.30 percent of GDP is employed, as well as businesses that play an important role to GDP in SMEs (Pongwirittorn & Udomoang, 2011), as well as restrictions on how to handle as large a business as possible. The important thing for SMEs is the problem of competition that enables them to survive in the short and long term, especially within the business (Pongwirittorn & Udomoang, 2011). From the literature review. Creating a competitive advantage that has been widely accepted by 450

academics has proven in theory that competitive advantage can be built on the ability and internal resources of a business. (Glavas & Mish, 2015), which is considered to be the key to a competitive advantage (Safarzadeh et al., 2015., Wang, 2014). Based on literature review, factors related to competitive advantage are: (1) Intellectual capital is an important factor for success in gaining competitive advantage (Kamukama, 2011), Consistent with the results of many studies. For example, research by Khalique & Hasson (2014), Papula & Volna (2014), Khalique et al. (2013), Jordon & Martos (2012), and Martin -de-Castro et al. (2011). This is because intellectual capital is the use of knowledge to create value added products and services to businesses. It is important for knowledge to be used economically. Businesses need to build intellectual capital to innovate from a knowledge base or from a knowledge base or from an intellectual capital. (Weerawardena & Mavondo, 2011). Especially for SMEs, intellectual capital is considered as an important asset to make a business successful. (Khalique et al. (2013), as well as Jardon & Martos (2012). Intelligence is very important as a source of competitive advantage in SMEs, enabling SMEs to survive in a competitive environment, and Khalique & Hassan (2014) says that intellectual capital is a factor. Most importantly, it affects competition in small and medium businesses. Intellectual capital also plays an important role in enhancing the performance of SMEs. (2) Innovation is what helps businesses survive in a rapidly changing environment (Klimas, 2014). The main factors that make the organization successful (Romero & Martine-Roman, 2012), especially small and medium businesses (Aini et al., 2013), have influenced business performance. It is therefore necessary to continuously innovate in order to accommodate the changes that occur (Klimas, 2014), and is a key factor in the competitive advantage (Yasin et al., 2014) in line with De Lara & Guimaraes (2014), which states that the influence of competitive advantage is driven by innovation, as Kamboj & Rahman (2014) says innovation has a direct impact on competitive advantage. And Lee & Hsieh (2010), which showed that innovation can have a direct influence on the competitiveness and sustainability. As with Hana (2013) study found that Innovation and support for innovation are critical to enabling businesses to achieve competitive advantage and Kamboj & Rahman (2014) found that innovation capabilities directly influence sustainable competitive advantage and (3) Dynamic capabilities From the literature review. Dynamic capacity is an important factor for creating competitive advantage (Corte & Gaudio, 2012). For example, Su et al. (2014), Li & Liu (2014), Schilke (2014), Lin & Wu (2014), Makkonen et al. (2014), Teece (2014), Cui & Jiao (2011), Jiao et al. (2010) and Wu (2010). Dynamic capabilities were presented as a new and fundamental concept. Competitive advantage (Teece, 2009), a concept that complements the absence of RBV, Each yet to identify opportunities and to adjust capacity in the future (Glavas & Mish, 2015) is the ability to modify resources with goals. Internal Resources and External Business Resources (Corte & Gaudio, 2012) By linking the capabilities of organizations embedded in business. Integration between internal and external business capabilities (Teece, 2014) in line with the changing external environment. By searching for opportunities and mix resources and knowledge. For a change to the new capabilities (Corte & Gaudio, 2012), it will affect the value 451

of the organization. This can lead to competitive advantage (Capron & Mitchell, 2009). The competitive advantage can be considered as a unique ability of a business that competitors cannot imitate, or if imitation may take time to imitate, such as the image of a business. Internal Management System (Tuan & Yoshi, 2010), based on the key factors of competitive advantage, is knowledge. Businesses can build on the experience, expertise, expertise of the people within the business. Then develop these into a source of business knowledge. The so-called. \"Intellectual Capital\" Businesses must draw on the knowledge or intellectual capital. The person has to develop a product or service to be unique, novel or to bring knowledge. This is the basis for innovation to the business, which will make the business is different from the competition. Innovation is the ability of a business to innovate. This is an important tool for businesses to gain competitive advantage. (Weerawardena & Mavondo, 2011), and there is a need for continuous innovation to enable businesses to survive in a rapidly changing environment (Hurley & Hult, 1998). Dynamic Capabilities and their ability to create and nurture dynamic capabilities. (Weerawardena & Mavondo, 2011) In addition, the advantage may be due to certain assets or resources that businesses already have (De Lara & Neves Guimaraes, 2014). So this research objective to study investigate antecedent the factors influencing the competitive advantage and explore influence of competitive advantage upon performance. This will enable us to understand the factors that affect the competitive advantage by having good performance of SMEs in Thailand. This will lead to the value creation for the enterprise allowing enterprises to compete and survive under current circumstances. 2. Methods 2.1 The Subject of the Study This study is quantitative research that provides broad, empirical empirical data that can be applied to all areas to check with the theoretical framework set forth by the researcher based on the principles, concepts and theories to get the findings on key issues. The researcher collected data using a questionnaire with business owners, managers, supervisors or related employees of SMEs in Buriram, Surin and Sisaket, Thailand, in 2014 from 128,516 sites (Office of Small and Medium Enterprises Promotion, 2017).We determine sample sizes that are suitable for data analysis with the LISREL program by using the Structural Equation Modeling (SEM) statistical technique. The sample size must be 20-10 for each variable in the research (Angsuchote et al., 2015). In this research, the researcher had variable observations in 16 models. So the sample size was appropriate and sufficient so it should have 320 (20 x 16). In addition, collecting data by mailing and meeting questionnaire at the establishment. In order to get the proportion of questionnaires responded to, the researcher collected 400 samples. Therefore, the sample size used in this study was 400 for each province totaling 1,200. The researcher used multistage random sampling method because the population in the study was large. It is important to select 452

the sample of the largest size. Then we selected sub-sample to the minor level and did this to the desired level (Kattiya & Suvajittanon, 2012). The population was divided into sub-groups in sequence. 2.2 The Tools of Research The questionnaire can be divided into 6 parts: 1) general information of respondents 2) about intellectual capital 3) about innovation 4) dynamic capabilities 5) About the competitive advantage and 6) about the performance. The creation of research tools; The researcher studied the related theoretical and literary concepts to define the operational definition and structure of the variables the researcher want to study. The researcher then created a questionnaire based on the operational definition that the developer of the instrumentation and the questionnaire has been improved to fit the research. And the researcher brings the questions that have been developed to the experts to examine the content validity of the questions from the study of related theoretical and literary concepts. Both domestic and foreign literature. When the expert examines the content of the questionnaire. The researcher modified the questionnaire to produce a draft questionnaire. After that, the researcher took the questionnaire to test validity by using the questionnaire developed by the researcher for 5experts to find the index of correspondence between the question and the objective (Index of Item Objective Congruence: IOC). The content validity of the questionnaire is 0.93, which is considered in the criteria (IOC>0.50) (Kanjanawasee, 2012). This shows that all questions in the questionnaire matched the questionnaire with the characteristics of the research objectives to be measured, content validity and suitability and cover the content that the researcher wants to study. It can be used to collect data. The reliability of the questionnaire was tested by 30 participants. This is not a research sample. The reliability of the questionnaire was 0.989. The reliability of the questionnaire and questionnaire with the α value of 0.70 and above was considered to be the confidence question (Kattiya & Suvajittanon, 2012). 2.3 Statistics Used in Data Analysis. The researcher has conducted statistical analysis that is appropriate and consistent with statistical data to meet the purpose of the research set. The statistics used to analyze the data are four parts. Part 1 Descriptive statistics. Used to describe or describe the attributes or properties of the distribution of variables. According to the characteristics of the group. Percentage, mean, and standard deviation were used to determine the basic statistics of the observed variables. Part 2 Statistical Analysis of Relationship Between Variables For the analysis of relationships between variables, Pearson's Product Moment Correlation Coefficient (Pearson's Product Moment Correlation Coefficient) is linear. Can identify the direction of the relationship. (Positive or negative) and the size of the relationship is at what level to serve as a basis for 453

analyzing causal factors and effects of dynamic capabilities of small and medium enterprises (Kattiya & Suvajittanon, 2012). Part 3 Statistical analysis of latent variables and mean values of variance were extracted. The researcher considered the reliability of the Construct Reliability (ρc) and the Average Variance Extracted (ρv) by using the formula (Diamantopoulos & Siguaw, 2000). Part 4 Statistical Analysis of Structural Equation Models The researchers used the analysis of structural equation modeling (SEM) to examine the harmony of the model with the empirical data (Model Fit). Appearance Model of Fit. The index used to check the fit of the model (Measurement Model) with the empirical data (Angsuchote et al., 2015). 3. Results The findings indicates that researchers divided the topic into five parts, with the details as follows. 1. General information of the respondents; Most respondents are male. 65.08% of them are 41-50 years old, 64.75% of them have experience of working 6-10 years, representing 37.58%. 34.08% of businesses have registered capital of not more than 1,000,000 baht. It is 64.00%. It has been in operation since the establishment of the business for 5 - 9 years, accounting for 25.83%. The number of employees is not more than 20 persons or 38.33%. 2. Average Data of Variables; Intellectual Capital, Innovation, Dynamic Capabilities, Competitive advantage and performance. 2.1 Intellectual Capabilities; The respondents' level of opinion on intellectual capital was at a high level. When considering each aspect, it was found that the relationship capital. The highest level of opinion was the human capital. And structural capital, respectively. 2.2 Innovation; The respondents' level of opinion about innovation was at a high level. When considering each aspect, it was found that corporate innovation was at the highest level of opinion followed by product innovation. Process innovation and the innovation side of the marketing, respectively. 2.3 Dynamic Capabilities; The respondents had a high level of feedback on overall dynamic performance. When considering each side, it found that the opportunities seizing capabilities. The highest level of feedback was the opportunity sensing capabilities. And the resource reconfiguration capabilities, respectively. 2.4 Competitive Advantage; The respondents had a high level of opinions on overall competitive advantage. When considering each aspect, it was found that the flexibility side had the highest opinion level, followed by the cost. And quality aspects respectively. 2.5 Performance; The respondents had a very high level of feedback on overall performance. When considering each aspect, it was found that the strategy had the 454

highest level of finance, followed by strategies and marketing, respectively. 3. From the results of the research model consistency check with empirical data, the first model analysis showed that the harmonic index was not consistent with the empirical data. Or does not meet the criteria set. Some important stats. Not yet meet the criteria set. The researcher then proceeded to modify the model by adjusting the parameters by agreeing to relax the initial agreement for the relative error. For Analytical results, after the model was modified, the model was found to be in harmony with the empirical data, with the six harmony indexes that met the acceptance criteria. The index values χ2 / df = 1.489, CFI = 0.999, GFI = 0.95, AGFI = 0.93, RMSEA = 0.047 and SRMR = 0.021. In harmony with the empirical data. 4. Route Analysis Results 4.1 Intellectual Capital (IC) has a direct positive influence on the Competitive Advantage of the business (CA) with a direct magnitude of 0.18 which is statistically significant at .05 level. 4.2 Innovation (IN) has a direct positive influence on the18 Competitive Advantage of the business (CA), with a direct magnitude of 0.61 which is statistically significant at .05 level. 4.3 Dynamic Capabilities (DC) has a direct positive influence on the Competitive Advantage of the business (CA), with a direct magnitude of 0.09 which is statistically significant at .01 level. 4.4 The Competitive Advantage of the business (CA) has a direct positive influence on the performance (PER) with a direct magnitude of 0.11, which is statistically significant at .01 level. 4.5 Intellectual Capital (IC), Innovation (IN) and Dynamic Capabilities (DC) have a positive indirect influence on Performance (PER) through Competitive Advantage (CA). The mean was 0.66, 0.13 and 0.09, respectively, which was statistically significant at .01 level. IC 0.66 PER 0.18 CA 0.11 0.61 0.13 IN 0.09 0.09 DC Fig 1: Model of Competitive Advantage 455

Acknowledgments and Legal Responsibility Thank you Rajamankala University of Technology Isan for budgeting for this research. 4. References Aini, E.K., Long Shen, D.C., Musadieq, M.A., &Hanayani, S. R. (2013). The Role of Innovation Capability on Business Performance at Small Medium Enterprises .Journal Profit, 7 (1), pp. 101-110. Angsuchote, S., Vijitwanna, S. &Phinyopanuwat, R. (2015). Stratistical Analysis for Social Science Research and Behavioral: Techniques for Using LISREL (Edition 40). Bangkok: Rongphim Charoendi Mankhong Kanphim. Boone, L.E. & Kurtz, D.L. (2010). Contemporary Business. (13th ed.). New York: John Wiley & Sons. Capron, L .& Mitchell, W. (2009). Selection capability :How capability gaps and internal social frictions affect internal and external strategic renewal .Organization Science, 20 (2), pp. 294-312. Corte, V.D., & Gaudio, G.D. (2012). Dynamic Capabilities : A Still Unexplored Issue with Growing Complexity .Corporate Ownership & Control, 9(4), pp. 327-339. Cui, Y., & Jiao, H. (2011). Dynamic Capabilities, Strategic Stakeholder Alliances and Sustainable Competitive Advantage :Evidence from China .The International Journal of Business in Society, 11(4), pp. 386-398. De Lara, F.F .& Neves Guimaraes, M. R. R. (2014). Competitive Priorities and Innovation in SMEs :A Brazil Multi-Case Study .Journal of Technology Management & Innovation, 9(3), pp. 51-64. Diamantopoulos, A .&Siguaw, A.D. (2000). Introducing LISREL :A Guide for the Uninitiated. London: Sage Publications. Glavas, A .& Mish, J. (2015). Resources and Capabilities of Triple Bottom Line Firm :Going Over Old or Breaking New Ground? Journal Business Ethics, 127, pp. 623-642. Hana, U. (2013). Competitive Advantage Achievement through Innovation and Knowledge. Journal of Competitiveness, 5(1), pp. 82-96. Hurley, R. F. & Hult, G.T.M. (1998). Innovation, Market Orientation and Organizational Learning :An Integration and Empirical Examination .Journal of Marketing, 62, pp. 42-54. Jardon, C.M. & Martos, M. S. (2012). Intellectual Capital as Competitive Advantage in Emerging Clusters in Latin America .Journal of Intellectual Capital, 13 (4), pp. 462-481. Jiao, H., Wei, J., & Cui, Y. (2010). An Empirical Study on Paths to Develop Dynamic Capabilities: from the Perspectives of Entrepreneurial Orientation and Organizational Learning. Frontiers of Business Research in China, 4(1), 47-72. Kamboj, S .& Rahman, Z. (2014). Marketing Capability, Innovative Capability and Sustainable Competitive Advantage :A Conceptual Framework .Research and Sustainable Business, pp. 711-717. 456

Kamukama, N., Ahiauzu, A., & Ntayi, J.M . (2011) .Competitive Advantage :Mediator of Intellectual Capital and Performance .Journal of Intellectual Capital, (1)12, pp.164-152 . Kanjanawasee, S .(2012) .Multi-level Analysis (Edition 5 ngkok :Chulalongkorn University Printing House . Kattiya, S &Suvajittanon, W .(2012) .Research and Statistic Models, Bangkok :Prayoonwong Printing Co .LTD. Khalique, M .& Hassan, A .(2014) .Intellectual Capital in SMEs Operating in Boutique Sector in Kuching, Malaysia. The IUP Journal of Management Research, (2)8, pp. 17-28. Khalique, M., Isa, A.H.B.M., & Shaari, N.B. (2013). Predicting the Impact of Intellectual Capital Management on the Performance of SMEs in Electronics Industry in Kuching, Sarawak . The IUP Journal of Knowledge Management, 11(4), pp. 53-61. Klimas, P. (2014). How Objective Are Subjective Measures of Organizational Innovativeness, Really? Management and Business Administration, Central Europe, 22(2), pp. 30-44. Lee, J.S. & Hsieh, C.J. (2010). A Research in Relating Entrepreneurship, Marketing Capability, Innovative Capability and Sustained Competitive Advantage. Journal of Business & Economics Research, 8(9), pp. 109-119. Li, D., & Liu, J. (2014). Dynamic capabilities, environmental dynamism, and competitive advantage :Evidence from China .Journal of Business Research, 67, pp. 2799-2793. Lin, Y. & Wu, L.Y. (2014). Exploring the Role of Dynamic Capabilities in Firm Performance under the Resource-Based View Framework. Journal of Business Research, 67(3), pp. 407-413. Makkonen, H., Pohjola, M., Olkkonen, R., & Koponen, A. (2014). Dynamic Capabilities and Firm Performance in a Financial Crisis. Journal of Business Research, 67(1), pp. 2707-2719. Martin-de-Castro, G., Delgado-Verde, M., Lopez-Saez, P. & Navas-Lopez, J.E. (2011), Towards An Intellectual Capital-Based View of the Firm: Origins and Nature .Journal of Business Ethics, 98, pp .649-662. Office of Small and Medium Enterprises Promotion. (2017). White Paper on Small and Medium Enterprises of Thailand in 2016. Bangkok :Thang Sa-Duak (Thai) Co. LTD. Papula, J. & Volna, J. (2014). The Level of Intellectual Capital Management in Slovak Companies .Faculty of Management, Comenius university, Braislava, Slovak Republic, pp. 135-144. Pongwirittorn, R. & Udomoang, S. (2011). The Management of Small and Medium Enterprises to Achieve Competitive Advantages Northern of Thailand. Journal of Management Sciences. 28 (1), pp. 49-63. Romero, I. & Martine-Roman, J.A. (2012). Self-Employment and Innovation. Exploring the Determinants of Innovative Behavior in Small Business. Research Policy, 41(1), pp. 178-189. http://dx.doi.org/10.1016/j.respol.2011.07.005 Safarzadeh, T., Dahghan, E., Pazireh, M., & Pouraskari, F. (2015). Checking the Relationship 457

between Strategic Leadership, Competitive Advantage and Organizational Performance with Mediating Role of Innovation. International Journal of Business and Behavioral Sciences, 5(1), pp .11-18. Schilke, O. (2014). No The Contingent Value of Dynamic Capabilities for Competitive Advantage :The Nonlinear Moderating Effect of Environmental Dynamism .Strategic Management Journal, 35, pp. 179-203. Su, H.C., Linderman, K., Schroeder, R.G., & Van de Vaen, A.H. (2014). A Comparative Case Study of Sustaining Quality as a Competitive Advantage. Journal of Operations Management, 32, pp. 429-445. Teece, D.J. (2014). Dynamic Capabilities-based Entrepreneurial Theory of the Multinational Enterprise .Journal of International Business Studies, 45, pp. 8-37. ________. (2009). Dynamic Capabilities and Strategic Management. Oxford University Press, Oxford, UK. Tuan, N.P.& Yoshi, T. (2010). Organisational Capabilities, Competitive Advantage and Performance in Supporting Industries in Vietnam .Asian Academy of Management Journal, 15(1), pp. 1-21. Wang, C.H. (2014). How Relational Capital Mediates the Effect of Corporate Reputation on Competitive Advantage :Evidence from Taiwan High-Tech Industry. Technological Forecasting & Social Change, 82, pp. 167-176. Weerawardena, J. & Mavondo, F.T. (2011). Capabilities, Innovation and Competitive Advantage. Industrial Marketing Management, 40, pp. 1220-1223. Wu, L.Y. (2010). Applicability of the Resource-Based and Dynamic-Capability Views under Environmental Volatility. Journal of Business Research, 63(1), pp. 27-31. Yasin, G., Nawab,S., Bhatti, K., & Nazir, T. (2014). Relationship of Intellectual Stimulation, Innovations and SMEs Performance :Transformational Leadership a Source of Competitive Advantage in SMEs .Middle-East Journal of Scientific Research, 19(1), pp. 74-81. 458

ISSSM-0396 An Exploratory Analysis of Consumers’ Selection Factors on Green Hotel in Taiwan Chien-Hui Honga, Hung-Ming Linb Department of Business Administration, Minghsin University of Science and Technology, Taiwan E-mail: [email protected] a, [email protected] 1. Background Starting from 2009, the environmental protection Administration of Taiwan began to give the green hotel mark to the hotels which fulfill the requirement. In 2012, the green hotel mark was revised and became three levels, the gold, the silver and the bronze. Until now, only seven hotels in this island obtain the green hotel mark. Green hotels are the environmentally-friendly hotels which are offering their efforts in saving water, saving energy, recycling, reducing solid waste, green purchasing,…etc. to protect environment of our mother land. It is our concern that what kind of consumers will choose the green hotel while they are travelling? They are willing to stay in a green hotel for what kinds of reasons. The purpose of this study includes (1) to find out how consumers’ awareness of the green hotels. (2) to find out consumers’ opinions regarding to environmental protection policies providing by hotel. (3) to find out the influential factors that affect consumers’ willingness to choose green hotel. (4) making some suggestions to hotel owners and managers and also to the government for promoting more green hotels in Taiwan. It is a worldwide trend to become a green hotel, not only to help environment but also to raise the hotel’s reputation. The hotel industry in Taiwan should put more efforts on green services. 2. Methods A questionnaire survey was conducted, and 400 copies questionnaire were collected from the customers at two international travel exhibitions in Taiwan. 394 samples were valid for this research. Using SPSS statistic analysis software, this research conducts narrative statistical analysis, factor analysis, reliability analysis and variance analysis. There are 15 questions in the questionnaire. After factor analysis, we choose four main factors to be used in variance analysis. The four main factors are hotel facilities, hotel environment, rooms /housing and the last one is the cost of staying. By variance analysis we discovered the relationship between person basic characteristics and the four factors. 3. Results The results of this research are following: (1) Consumers’ awareness of green hotels is very high. Most of the people are willing to follow the environmental protection policies. But the price of room or lodge is still a major concern. Usually they will struggle between environmental 459

protection and saving the money. (2) After variance analysis, we find out gender and education levels do not affect the preference of choosing green hotels. But age has high significant influence in cost and rooms/housing. Marriage has high significant influence in hotel environment factor. Occupation has high significant influence in hotel environment and rooms/housing factors. (4) There are some suggestions made to hotel owners and managers. Some promoting policies can be used to encourage consumers to come. For example, giving discount to whom bring their own washing utensils. Seeking financial subsidy from central or local government is another way to improve hotel’s environment. (5) Government should play a bigger role in promoting green hotels. Besides educating people the importance of environmental protection, government can offer more knowledge to hotel industry about the necessary and requirement of green hotel mark. Government also can provide financial aids to the hotels which have difficulties to buy expensive facilities for green services but are willing to do some contribution to environmental protection. Keywords: Green Hotel, Green Hotel Mark, Factor Analysis, Variance Analysis 4. References 1. Hsieh, C.-L. (2006). A study of the effects of consumers’ green involvement on green hotels’benefit ─ An empirical study on Han-Hsien International Hotel (Unpublished master dissertation). National Kaohsiung University of Hospitality and Tourism, Taiwan. 2. Yang, Y.-T. (2012). Determinants of acceptances of green mark for eastern hotels in Taiwan (Unpublished master dissertation). National Dong Hwa University, Taiwan. 3. Green Hotel Association, 2018. What are green hotels? Retrieved from http://www.greenhotels.com。 4. Environmental Protection Administration, green living network, 2018. Retrieved from http://greenliving.epa.gov.tw。 460

ISSSM-0273 The MQTT-Based Topic Naming Schemes for Smart Care Cheng-Min Lin Department of Digital Living Innovation, Nan Kai University of Technology, R.O.C Gerontechnology Research Center, Nan Kai University of Technology, R.O.C E-mail: [email protected] Abstract With the growth of an aging society, elderly care is a crucial concern. Home care, community care, and institutional care require huge resources and manpower. Therefore, smart care has become very essential. This study presents three naming schemes to facilitate Internet of Things (IoT)-based smart care for the protocol of Message Queuing Telemetry Transport (MQTT): an ObjectiveLocationOwner naming scheme, an identity naming scheme, and a multibroker naming scheme. The latter two proposed novel topic naming methods have the advantage of reducing network traffic and maintaining the benefits of lightweight protocols. Keywords: IoTs, MQTT, Smart Care, Elderly Care 1. Background The aging society is growing. Elderly care concerns are becoming increasingly important. Elderly care is generally divided into three categories, namely home care, community care, and institutional care. Because elderly care requires huge resources and manpower, smart care becomes crucial. Scholars have proposed numerous methods and applications, such as human activity recognition systems (Jalal, Kamal, and Kim, 2014), remote health monitoring systems (Guan, Shao, and Wu, 2017), and smart care beds (Hong, 2018). The development of Internet of Things (IoT) is the foundation for smart care. Because of limited resources, an extremely lightweight transport protocol is necessary. Message Queuing Telemetry Transport (MQTT) is a lightweight transport protocol frequently used for IoT. MQTT is a publish-subscribe-based messaging protocol including clients and a server. Clients can be categorized as publishers and subscribers, whereas some clients are both. The subscriber subscribes to a topic and then receives content related to a topic from a broker when a publisher publishes the topic with its content to the broker. Although MQTT is a widely used protocol in several systems worldwide, no MQTT-Topic naming standard is available for the open data of smart cities (Tantitharanukul, Osathanunkul, Hantrakul, Pramokchon, & Khoenkaw, 2016). This paper presents two naming schemes to develop an IoT-based smart care based on the ObjectiveLocationOwner naming scheme proposed by Tantitharanukul et. al (Tantitharanukul, 461

2016). The two schemes are based on elderly care and primarily consider the process of naming the topic of IoT. This paper presents two novel topic naming methods that have the advantage of reducing network traffic and maintaining the benefits of lightweight protocols. The remainder of this paper is organized as follows: Section 2 presents MQTT techniques. In Section 3, three naming schemes for MQTT are presented. Section 4 discusses results and compares the schemes. Section 5 provides conclusions. 2. Message Queuing Telemetry Transport The primary concept applied in IoT is that everything can be connected to the Internet. Everything comprises numerous distributed systems. A distributed system includes resources and computed tasks. These systems may come from different manufacturers. The exchange of information among these systems is difficult. IoT protocols are used to solve this problem. MQTT is one of the IoT protocols provided by the Organization for the Advancement of Structured Information Standards (OASIS). It is used to interconnect devices or systems. MQTT is a lightweight messaging protocol that uses publish–subscribe technology. Therefore, MQTT is suitable for mobile services. MQTT is bandwidth efficient and conserves power. Fig. 1 shows the architecture of this protocol that uses the publish–subscribe mechanism. This protocol is different from HyperText Transfer Protocol in terms of its request or response mechanism. The architecture consists of two types of nodes. One node is brokers, whereas the other is clients, such as elderly 1, elderly 2, monitoring center, and ambulance in Fig. 1. These clients are smart devices worn by elderly people or installed in the monitoring center or ambulance. Fig. 1: MQTT Publish/Subscribe architecture For the monitoring center in Fig. 1, two topics including “fall down” and “go out” are subscribed 462

to actively receive information regarding events related to the client regarding the subscribed topic. For the ambulance, only one topic of “fall down” is subscribed. Therefore, an MQTT broker publishes the message of “go out/elderly 2” when it receives the message of “go out/elderly 2.” If the MQTT broker receives “fall down/elderly 1”, it informs the monitoring center and ambulance. According to OASIS Standard (OASIS, 2014), the operation of MQTT includes an exchange of a series of MQTT control packets between the broker and clients. An MQTT control packet comprises up to three parts (Fig. 2), namely fixed header, variable header, and payload. Fig. 2: Structure of an MQTT control packet Each MQTT control packet includes a fixed header (Fig. 3). In the first byte, bits 4–7 represent the type of the MQTT control packet. For example, the values of 3 and 8 are PUBLISH and SUBSCRIBE, respectively. Remaining bits 0–3 of the first byte in the fixed header include flags, which are specific to each MQTT control packet. The remaining length is up to 4 bytes. Therefore, an application is allowed to send control packets of size up to 268,435,455 (256 MB). Fig. 3: Fixed header format 3. Methods Although a long topic name can convey a complete message, the topic name has an upper length limit of 32,767 characters. This is a trade-off problem. Although the topic name is sufficiently long to convey complete information, longer transmission time is required. The original advantage of MQTT is thus nullified. Few studies are available on the practical application of topic naming, and only Tantitharanukul et. al. proposed a topic name for smart cities (Tantitharanukul, 2016). The detailed description of different methods is provided in the following sections. 463

3.1 ObjectiveLocationOwner Naming Scheme At least three things must be considered when installing any device in a smart city, namely an objective, a location, and an owner. Therefore, minimum three attributes, namely the objective, location, and owner, are presented in MQTT (Tantitharanukul, 2016). Tantitharanukul et. al proposed three components of the objective attribute, namely a primary objective, subobjective, and user-defined objective (Tantitharanukul, 2016). The indicators of international standard ISO 37120 are used as a guideline for the primary objective and subobjective. Publishers are allowed to define their user-defined objective. Tantitharanukul et. al proposed three components of location, which were attributed similar to the objective: country, province, and user-defined location. The country attribute is described using three roman characters represented by the alpha-3 code of the ISO 3166 standard. The second attribute is a province attribute. The user-defined location is defined by the publisher, similar to the user-defined objectives. The owner attribute is defined by the publishers. The publisher can decide the topic naming rules for the owner. In general, the attribute is one component. Moreover, avoiding the reuse of the manufacturer name is appropriate. Tantitharanukul et. al did not name this method. In this paper, the method proposed by Tantitharanukul et. al is termed the ObjectiveLocationOwner naming scheme. This method is advantageous because it simply and clearly describes the naming rules; however, the limitation of this method is the excessively long subject name, which counters the advantage of MQTT. Thus, this method is not suitable for mobile computing and large IoT systems. 3.2 ID Naming Scheme The advantage of the ObjectiveLocationOwner naming scheme is that it can understand its meaning from the name of the topic; however, the topic name is considerably long, thereby countering the advantage of the lightweight agreement of MQTT. In this paper, the identity (ID) naming scheme is used to replace the ObjectiveLocationOwner naming scheme. The publisher sends a message to the broker on the network, and the subscriber subscribes to the message from the broker by transmitting the identity instead of a large string of characters. The advantage of this method is that it can avoid a large amount of data transmission on the network, whereas its disadvantage is that a converted component is implemented in the broker. It is transparent to the conversion mechanism of identity and name for topics. Users employ the original method to publish and subscribe to topics. Whether it is publishers to brokers or brokers to subscribers, a conversion program is embedded in brokers, publishers, and subscribers (Fig. 464

4). Fig. 4: ID Naming Scheme 3.3 Multibroker Naming Scheme MQTT comprises clients and a broker. IoT based on MQTT has a star-shaped architecture. All subscribers and publishers exchange messages with the broker. Therefore, this architecture is not suitable for large IoT. When the number of subscribers or publishers is considerably large, the broker becomes inefficient because of the burden. MQTT does not provide information exchange among brokers. Therefore, a middleware was implemented in the broker. The middleware can be used to use the subscription and publishing mechanisms to connect several brokers. In Section 2-1, the ObjectiveLocationOwner naming scheme was introduced. For the objective or location naming scheme, three levels are available. Therefore, a three-level architecture of brokers can be implemented for the objective or location; however, having an objective and location simultaneously is not allowed. Only one of them can be used to implement a three-level architecture of brokers. 465

Fig. 5: A three-level architecture of brokers. 4. Results This paper presents three topic naming methods for smart care using the MQTT protocol. Table 1 compares the three methods. The ObjectiveLocationOwner naming scheme is provided by Tantitharanukul et. al in 2016. It provides a clear topic name; however, the topic name is too lengthy, which counters this advantage of MQTT. Both ID naming scheme and multibroker naming scheme require additional software for support. These methods implement lightweight protocols. The multibroker naming scheme can operate independently even if the broker in the cloud is disconnected. Table 1: Comparison of three methods Advantage Disadvantage ObjectiveLocationOwner naming scheme No extra software required The topic name is the longest ID naming scheme The topic name is the shortest Extra software required Multibroker naming scheme The system is more robust Extra software required Multiple brokers required 5. Conclusions Smart care for the elderly people must be based on IoT. The data is collected for analysis through the IoT. The length of the topic name is crucial because it affects system performance. In this paper, three naming schemes are presented, and their disadvantages and advantages are discussed. In the future, the effectiveness of the three methods will be thoroughly analyzed, and the use of IoT in smart care will be studied. 466

6. Acknowledgments The authors would like to thank the Ministry of Science and Technology of the Republic of China, Taiwan for financially supporting this research under Contract No. MOST 107-2637-E-252-001. 7. References Guan, K., Shao, M., and Wu, S. (2017). A Remote Health Monitoring System for the Elderly Based on Smart Home Gateway. Hindawi Wireless Communications and Mobile Computing, pp. 1-9. https://doi.org/10.1155/2017/5843504 Hong, Y. S.(2018). Smart Care Beds for Elderly Patients with Impaired Mobility. Hindawi Wireless Communications and Mobile Computing, pp. 1-12 https://doi.org/10.1155/2018/1780904 Jalal, A., Kamal, S. and Kim, D. (2014). A Depth Video Sensor-Based Life-Logging Human Activity Recognition System for Elderly Care in Smart Indoor Environments. Sensor, 14, pp. 11735-11759; doi:10.3390/s140711735 OASIS(2014). MQTT Version 3.1.1 OASIS Standard. Retrieved from http://docs.oasis-open.org/mqtt/mqtt/v3.1.1/os/mqtt-v3.1.1-os.html Tantitharanukul, N., Osathanunkul, K., Hantrakul, K., Pramokchon, P., & Khoenkaw, P. (2016). MQTT-Topic Naming Criteria of Open Data for Smart Cities. 2016 International Computer Science and Engineering Conference. IEEE. doi:0.1109/ICSEC.2016.7859892 467

ISSSM-0302 Factors Associated with Fall Risk Behaviors in Hospitalized Patients Akiko Hiyama School of Nursing, Sapporo City University, Japan E-mail: [email protected] Abstract Background: Falls are an important public health issue in communities, nursing homes, and hospitals. Although some studies have suggested risk factors that were associated with fall in hospitals, factors affecting fall risk behaviors remain unclear. The aim of this study was to investigate the factors related to fall risk behaviors. Methods: A retrospective design was used to examine 380 patients admitted to the general wards from September 2013 to August 2014. Logistic regression analysis was used to estimate fall risk factors that were associated with fall risk behaviors Results: Logistic models identified fall risk behavior: “moving with a lack of attention to safety, due to poor concentration” was associated with diminished attention (OR=4.8), overestimation (OR=4.3), history of fall (OR=2.8), wheelchair, walker, stick use (OR=2.4). “starting an action without checking the safety of the surroundings, due to hearing and/or vision impairment” was associated with hearing impairment (OR=3.7), diminished attention (OR=2.7), wheelchair, walker, stick use (OR=1.9), requiring assistance to move (OR=1.8). Conclusions: These results indicate that fall risk behavior was attributed to the stability of movement and diminished attention. It was suggested that a strategy that causes the patients to pay attention should be developed to prevent the fall risk behavior. Keywords: fall risk, risk behavior, risk factor 1. Background Falls are an important public health issue in communities, nursing homes, and hospitals. In Japan, falls account for up to 19.3% of overall accidents, and 10% of fallers experience a serious injury or death (Japan Council for Quality Health Care, 2015). Falls may lead to a change in the life style, in that some people who have fallen become afraid of falling. Fear of falling may cause gait changes, activity restriction and deconditioning (Vellas et al., 1997), and falls associated with worsening of gait function (Pua et al., 2017). Therefore, it is important to understand what factors may lead to falls. 468

Some studies have suggested risk factors that were associated with falls in hospitals. Fall-related factors include a history of falling, reduced vision, unsteady gait, poor balance, altered mental status, cognitive impairment, decreased functional ability, chronic disease, poor lighting, condition of the ground surface, type of foot covering, inadequate assistive devices, structural design of bathrooms and grab bars, design of furniture, and improper use of assistive devices (Payson & Haviley, 2007; Cox et al., 2015; Oliveira et al., 2017). The problem appears to lie in the fact that the fall risk factors target elderly patients, and do not clarify behaviors directly related to the falls. Although several studies have focused on fall risk behavior, the subjects included people in the community (Clemson et al.,2003; Toba et al., 2005), and fall risk behavior in hospital has not been targeted. The aim of this study was to investigate the factors related to fall risk behaviors. 2. Methods 2.1 Design and Setting In order to determine the relationship between fall risk behaviors and risk factors of falling, this study used a descriptive, correlational, retrospective design. This study was conducted at four hospitals in Sapporo City, Hokkaido, Japan. The four hospitals were teaching hospitals with 80 to 312 beds, including general beds. 2.2 Sample and Data Collection The case sample consisted of 380 patients from 20 generic wards at 4 hospitals in Japan. Patients admitted to the hospitals over the study period (from September 2013 to August 2014) were enrolled. Inpatient administrative records, fall risk assessment results, and fall incident reports were combined as the data for analysis. Ethical approval was granted by the Institutional Review Board of the researcher’s institute. The data collection forms were anonymized, with no reference to names or identification numbers. 2.3 Instruments The instrument used for data collection was organized to collect 2 types of data: (1) general and medical information and (2) fall-related information. General information included gender and age. Medical information included the diagnosis and length of the hospital stay. The number of falls during admission, fall risk behavior, and the fall risk assessment tool output were included as the fall-related information. The fall risk behaviors were analyzed quantitatively (Hiyama, 2017) based upon the behaviors of the inpatients who had fallen. The predictive model of fall using the fall risk behavior had good sensitivity (84.0%, 95%CI: 0.75-0.93) for the prediction of fall (Hiyama, 2018). In this study, the fall risk behaviors: “moving with a lack of attention to safety, due to poor concentration” and 469

“starting an action without checking the safety of the surroundings” were observed. The fall risk assessment tool (JNA, 2003, P5) that is commonly used in Japanese hospitals was developed based upon 40 fall risk factors: (1) history of disease, (2) perception (3 items; balance disturbance /visual impairment/ hearing impairment), (3) motor function (4 items; muscle weakness of leg/ paralysis/ numbness/ osteoarthropathy), (4) mobility (5 items; staggering, wheelchair, walker, stick use/ not requiring help with walk/ requiring assistance to move/ bedridden), (5) cognition (4 items; dementia/ restlessness/ cognitive dysfunction/ consciousness disorder/ delirium), (6) medication (5 items; hypnotic medication/ analgesic medication/ narcotic medication/ purgative medicine/ hypotensive diuretic), (7) elimination (8 items; incontinence/ frequent urination/ toilet is far from the bed/ toileting at night/ use of a portable toilet/ use of a toilet for wheelchairs/ use of an indwelling catheter/ requiring assistance for toileting), (8) treatment stage (5 items; fever/ anemia/ surgery within three days/ rehabilitation/ worsening of the medical condition), (9) personality (5 items; moving without use the nurse call/ unable to nurse call/ diminished attention/ overestimation/ unadapted to the environment). The items from the fall risk assessment tool were observed as risk factors of fall. Patients were assigned binary scores, 0 or 1, by the research nurses. 2.4 Data Analysis SPSS 22.0 software (IBM Corp, U.S.) was used for the data analysis. Continuous data were reported as the mean and standard deviation (SD) or median and interquartile range, Mann-Whitney Test was applied for comparison between groups by the fall risk behavior. Categorical data were examined using frequency distribution, Chi square test and Fisher’s exact test to detect independence between fall risk behaviors and risk factors of fall. The fall risk behavior was examined using logistic regression. Logistic regression, after adjustment for age, was used to estimate fall risk factors that were associated with fall risk behaviors, providing an odds ratio (OR) with corresponding 95% confidence interval (CI). Statistical significance was set at a P-value of <0.05, 2-tailed. 3. Results The participants were a mean age of 68.9 (SD20.0) years with 187(49.2%) males and 193(50.8%) females. The length of stay was 18.6 (SD31.9) days. Forty-two (11.1%) were reported to have been “moving with a lack of attention to safety” of which 18(42.9%) had fallen “due to poor concentration”, 96 (25.3%) were reported to have been “moving with a lack of attention to safety, due to poor concentration” of which 35 (36.5%) had fallen. The median age of those who had been “moving with a lack of attention to safety” was 76 years (Interquartile range=65-84 years), and the median age those who did not was 74 years (Interquartile range=6-30 years); there was no significant difference. The median length of stay of those who had been “moving with a lack of attention to safety” was 6 days (Interquartile 470

range=6-30 days), and the median age those who did not was 14 (Interquartile range=1-23); p<0.05. The median age of those who had been “starting an action without checking the safety of the surroundings” was 81 years (Interquartile range=73-86.5 years), and the median age those who did not correspond was 71 (Interquartile range=59-81); p<0.05. The median length of stay of those who corresponded “starting an action without checking the safety of the surroundings” was 15 days (Interquartile range=4-31.5 days), and the median age of those who did not was 14 (Interquartile range=1-19.75); p<0.05. Table 1 shows the independence between fall risk behaviors and risk factors of fall. Stepwise regression analysis suggested four significant risk factors of fall (Table 2, 3): “moving with a lack of attention to safety, due to poor concentration” was associated with diminished attention (OR=4.8), overestimation (OR=4.3), history of fall (OR=2.8), wheelchair, walker, stick use (OR=2.4). The risk factor: “starting an action without checking the safety of the surroundings, due to hearing and/or vision impairment” was associated with hearing impairment (OR=3.7), diminished attention (OR=2.7), wheelchair, walker, stick use (OR=1.9), requiring assistance to move (OR=1.8). 4. Discussion It is well-known that declining motor function leads to the risk of fall, and this discussion focuses on the psychological perspective. Logistic regression, after adjustment for age indicated “moving with a lack of attention to safety, due to poor concentration” was associated with diminished attention, overestimation, history of fall, wheelchair, walker, stick use. A lack of attention is related to personal psychological tendencies as diminished attention, overestimation, history of fall, which was related to the individual's psychological tendency of the condition of the disease. The mistake of the self-recognition limit posture indicates the difference between the maximum distance of recognizing that attitude control is possible and the distance in actual action, and has been related to falls several times (Takatori, Shomoto, & Shimada, 2009), and overestimation is considered to be a perceptual error of activity ability. The decision making of \"maybe all right\" is similar to a mistake that is a vacuity or failure in judgment and/or reasoning, which are carried out in concretely deciding on the means for achieving objective selection or purpose (Reason, 1990/2014, P13), and overestimation can be regarded as a behavior due to a patient's human error. Further study should explore the relationship between fall risk behaviors and human error. Diminished attention was commonly associated with both fall risk behaviors. Elderly people who extended the center of gravity fluctuation trajectory while performing a double task, tended to exhibit decreased maximum muscle strength and walking speed (Nishimura & Naruse, 2012), and the walking time during a double task condition is a fall risk factor (Yamada, 2009). 471

Therefore, to achieve effective prevention, the goal should be to put the patient into an environment that simplifies living behavior and does not require them to perform multiple tasks. 5. Conclusion Logistic models identified that the fall risk behavior: “moving with a lack of attention to safety, due to poor concentration” was associated with diminished attention, overestimation, history of fall, wheelchair, walker, stick use. Furthermore, “starting an action without checking the safety of the surroundings, due to hearing and/or vision impairment” was associated with hearing impairment, diminished attention, wheelchair, walker, stick use, requiring assistance to move. These results indicate that fall risk behavior was attributed to the stability of movement and diminished attention. It was suggested that a strategy that causes the patients to pay attention should be developed in order to prevent the fall risk behavior. Acknowledgments This work was supported by the JSPS KAKENHI, Grant Number JP 15K20669. 6. References Clemson, L., Manor, D., & Fitzgerald, M. (2003). Behavioral Factors Contributing to Older Adults Falling in Public Places. Occupation, Participation and Health, 23(33), 107-117. Cox. J., Thomas-Hawkins, C., Pajarillo, E., DeGennaro, S., Cadmus, E., Martinez, M.. Factors associated with falls in hospitalized adult patients. Applied Nursing Research, 28(2), 78-82. doi: 10.1016/j.apnr.2014.12.003 de Oliveira, D. U., Ercole, F. F., de Melo, L. S., de Matos, S. S., Campos, C. C., Fonseca, E. A. M. (2017). Evaluation of falls in hospitalized elderly. Journal of Nursing UFPE on line, 11, 4589-4597. doi:10.5205/1981-8963-v11i11a231198p4589-4597-2017 Hiyama, A., Nakamura, K. (2017). Behaviors of Hospitalized Patients at High Risk of Fall. Japanese Journal of Nursing Research, 40(4), 657-665. Hiyama, A., Nakamura, K. (2018). Evaluation of inter-rater reliability and accuracy of the Fall Risk Behavior Assessment Tool (FRBA-Tool) for prediction of the risk of fall. Journal of Medical Safety, 2018, 22-28. Japan Council for Quality Health Care. The Project to Collect Medical Near-Miss/Adverse Event Information 2015 annual report. Retrieved from http://www.med-safe.jp/pdf/year_report_2015.pdf Japanese Nursing Association (2003). Japan Nursing white paper 2003. Tokyo: Japanese Nursing Association Publishing Company. Payson, C. A., & Haviley, C. A. (2007). Patient falls assessment and prevention Global Edition: Opus Communications, 5, Massachusetts: HCPro Inc. Pua, Y. H., Ong, P. H., Clark, R. A., Matcher, D. B., Lim, E.C. (2017). Falls efficacy, postural balance, and risk for falls in older adults with falls-related emergency department visits: 472

prospective cohort study. BMC geriatrics. 17;291. doi: 10.1186/s12877-017-0682-2 Nishimura, M., Naruse, K. (2012). Effect of cognitive task on the performance of elderly people. Journal of Sport Science in Nara Women’s University, 14, 37-43. Reason, J. T. (1990). Human Error. NY: Cambridge University Press. Takatori, K., Shomoto, K., & Shimada, T. (2009)．Relationship between Self-Perceived Postural Limits and Falls among Hospitalized Stroke Patients．Journal of Physical Therapy Science， 21(1)，29-35． Toba, K, Okochi, J., Takahashi, T., Matsubayashi, K., Nishinaga, M., Yamada, S., Takahashi, R., Nishijima, R., Kobayash, Y., Machida, A., Akishita, M. Sasaki, H. (2005). Development of a portable fall risk index for elderly people living in the community. Japanese Journal of Geriatrics, 42(3), 346-352. Vellas, B.J., Wayne, S.J., Romero, L. J., Baumgartner, R. N., Garry, P.J. (1997). Fear of falling and restriction of mobility in elderly fallers. Age Ageing, 26(3), 89-93. Yamada, M. (2009). Examination of the Preventive Effect of Attention Function Training on Falls for Community Dwelling Elderly: A Randomized Control Trial. Rigakuryoho kagaku, 24(1), 71-76. Table 1. Independence between fall risk behaviors and risk factors of fall 473

Starting an action without checking the Moving with a lack of attention to safety, safety of the surroundings, due to hearing due to poor concentration and/or vision impairment Total Yes (n=42) No (n=338) P Yes (n=96) No (n=284) P History of falls 120 26 94 0 46 74 0 Balance disturbance Visual impairment 10 2 8 0.31 5 5 0.13 Hearing impairment Muscle weakness of leg 57 5 52 0.65 23 34 0.008 Paralysis Numbness 40 5 35 0.79 24 16 0 Osteoarthropathy Stagger 145 20 125 0.24 54 91 0 Wheelchair, walker, stick use Not needs help with walk 15 1 14 0.49 6 9 0.224 Need assistance to move Bedridden 31 6 25 0.11 7 24 0.83 Dementia Restlessness 41 6 35 0.43 12 29 0.57 Cognitive dysfunction Consciousness disorder, delirium 87 14 73 0.12 39 48 0 Hypnotic medication Analgesic medication 158 27 131 0.003 61 97 0 Narcotic medication Purgative medicine 129 8 121 0.04 20 109 0.002 Hypotensive diuretic Incontinence 119 19 100 0.05 49 70 0 Frequent urination Toilet far from bed 39 3 36 0.56 6 33 0.17 Toileting at night Use portable toilet 63 12 51 0.05 26 37 0.002 Use toilet for wheelchair Use indwelling catheter 35 9 26 0.009 17 18 0.002 Need assistance for toileting Fever 92 19 73 0.002 40 52 0 Anemia Surgery within three days 32 7 25 0.07 18 14 0 Rehabilitation Becoming worse his medical condition 106 12 94 1 26 80 0.89 Moving without use nurse call Unable nurse call 73 0 65 15 58 0.37 Diminished attention Overestimation 13 1 12 1 3 10 1 Unadapted to environment 77 7 70 0.69 26 51 0.08 124 10 114 0.22 33 91 0.71 38 5 33 0.59 13 25 0.24 40 5 35 0.79 11 29 0.71 20 2 18 1 8 12 0.18 124 13 111 0.86 38 86 0.1 10 2 8 0.3 5 5 0.13 65 15 50 0.002 25 40 0.012 54 5 49 0.82 18 36 0.18 64 11 53 0.12 24 40 0.02 27 5 22 0.2 9 18 0.36 55 3 52 0.17 13 42 0.87 24 3 21 0.74 6 18 1 114 21 93 0.004 39 75 0.01 59 5 54 0.65 16 43 0.75 43 12 31 0.001 20 23 0.001 45 9 36 0.07 19 26 0.01 40 14 26 0 20 20 0 25 10 15 0 11 14 0.03 67 6 61 0.67 20 47 0.36 Table 2. Results of logistic regression model: Moving with a lack of attention to safety, due to poor concentration OR 95% CI p-value 0.01 History of falls 2.775 [ 1.31 , 5.87 ] 0.03 0.00 Wheelchair, walker, stick use 2.363 [ 1.101 , 5.07 ] 0.01 Diminished attention 4.789 [ 2.068 , 11.089 ] Overestimation 4.266 [ 1.556 , 11.69 ] Stepwise regression analysis using method of maximum likelihood,　Adjusted for risk factors of fall and age, OR means odds ratio, CI means confidence interval, n ＝380 Table 3. Results of logistic regression model: Starting an action without checking the safety of the surroundings, due to hearing and/or vision impairment 474

OR 95% CI p-value 0.00 Hearing impairment 3.691 [ 1.736 , 7.848 ] 0.04 0.05 Wheelchair, walker, stick use 1.855 [ 1.027 , 3.35 ] 0.02 Need assistance to move 1.814 [ 0.997 , 3.299 ] Diminished attention 2.653 [ 1.213 , 5.807 ] Stepwise regression analysis using method of maximum likelihood,　Adjusted for risk factors of fall and age, OR means odds ratio, CI means confidence interval, n＝380 475

ISSSM-0306 Establishing the Isolated Community Prediction Model Based on Machine Learning Theory and Spatial Information Yuanfang Tsaia, Meiting Liub The department of Social and Regional Development, National Taipei University of Education, Taiwan E-mail: [email protected] a, [email protected] 1. Background In recent years, the impact of global climate change increases significantly, and extreme rainfall occurs more frequently that led to flooding, landslides and debris flow in Taiwan has been increasing. The impact of short-duration heavy rainfall has caused many settlements to become \"Isolated Community\" due to the interruption of external traffic. Predicting whether or not Isolated disaster will occur not only keeping people’s life safety but also protects their property. 2. Methods This study uses the supervised learning algorithms of machine learning to establish a predictive model of the probability of becoming an Isolated in mountain villages, and selects the most suitable model. The Geographic Information System (GIS) is used to establish Isolated disaster spatial indicators, including topographic factors, human factors and hydrological factors, such as average slope, collapse, number of roads and maximum rainfall intensity of 22 kinds of factors. By using the independent sample T-test and the single-variable feature selection of machine learning select suitable features. 3. Results With the appropriate feature selection method constructed, it can reduce the sample training time and improve the accuracy of prediction. After eliminating 7 factors, four different models are used for training. The models are Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Decision Tree (DT) and Logistic Regression. Using the k-fold cross-validation technique to draw the learning curve and the verification curve to diagnose the learning algorithm, adjust the parameters appropriately, correct the model performance, calculate the performance index through the confusion matrix, and then use the receiver operating characteristic curve measuring the quality of the classifier. This study used the 2015 Sudler typhoon event to predict and validate the probability of each model. The conclusion is that the two modes of KNN and DT belong to the non-1 or 0 category analysis, but for the island event, the occurrence of the Isolated event has great variability, and the two modes of Logistic and SVM belong to the category analysis of the probability judgment, so it is more effective to capture the occurrence of an island incident. SVM has a higher probability of prediction than Logistic, so this study considers that 476

SVM mode has an optimal predictive value for Isolated prediction. Keywords: Isolated Effect, Extreme Rainfall, Supervised Learning, Machine Learning, Prediction Model 4. References 1. Cortes, C.,Vapnik, V. (1995).Support-Vector Networks.Machine Leaming.Kluwer Academic Publishers, Boston. Manufactured in The Netherlands,20, 273-297. 2. Voyant, C.,Notton, G.,Kalogirou, S.,Nivet, M.L.,Paoli, C., Motte, F., Fouilloy, A. (2017).Machine learning methods for solar radiation forecasting: A review.Renewable Energy ,105 (2017),569-582. 3. Alpaydın, E. (2004).Introduction to Machine Learning(Second Edition.).Massachusetts London, England﹕The MIT Press Cambridge. 4. Kavakiotis, I., Tsave, O., Salifoglou, A., Maglaveras N., Vlahavas I., Chouvarda I.(2017).Machine Learning and Data Mining Methods in Diabetes Research. Computational and Structural Biotechnology Journal ,15 (2017),104–116. 5. Quinlan, J.R. (1986).Induction of Decision Trees.Machine Learning 1: 81-106, 1986 Kluwer Academic Publishers, Boston - Manufactured in The Netherlands 6. Inman, R.H., Pedro, H.T.C., Coimbra, C.F.M. (2013). Solar forecasting methods for renewable energy integration, Prog. Energy Combust.Sci, 39 (2013),535~576. 7. Ravinesh,C.D., Şahin, M. (2015). Application of the extreme learning machine algorithm for the prediction of monthly Effective Drought Index in eastern Australia.Atmospheric Research,153 (2015), 512–525. 8. Aggarwal, S.K., Saini, L.M. (2014). Solar energy prediction using linear and non-linear regularization models: a study on AMS (American Meteorological Society)2013-14 solar energy prediction contest.Energy ,78 (2014),247~256. 9. Raschka, S. (2015).Python Machine Learning(1st edition.).published September 23rd 2015﹕Packt Publishing. 10. Sheng-Feng, Ting (2012). Study the concept of disaster risk and vulnerability analysis at Taitung County hillside fields (master's thesis, National DONG HWA University, Hualien). Retrieved from https://hdl.handle.net/11296/92nqu7 11. Kuei-Heng, Lu(2015). A High Efficiency Massive Spam Filtering Cloud System Based on Machine Learning (master's thesis, University of KANG NING, Tainan). Retrieved from https://hdl.handle.net/11296/j756v5 12. Chih-Chiang, Lee(2011). Community independent discussion on disaster-preparedness promotion –Mau forest Wanshan tribes of the city as an example (master's thesis, National Kaohsiung University of Science and Technology). Retrieved from https://hdl.handle.net/11296/34x2kx 477

13. Shue-Yeong, Chi, Yu-Wen, Lin, & Che-Wei, Shen.(2017) Construction Building Management Resume Database and Relationship Analysis of Crucial Hazard Factors in Hillside Residential Community(No. 10515G0013). Retrieved from http://www.abri.gov.tw/tw/research/show/2574/p/print 14. Jyun-Kai, Chen, Jhong-Jhih, Lin, Bo-Syun, Lin, Guo-Shih Shao, & Shun-Min, Wang.(2014) Risk Management of slope disaster. Sinotech Engineering, 125, 45-55. 15. Yi-Jyun, Lin(2009). A Study of Machine Learning Approach for Analyzing Personal Handwriting Features and Identity Identification (master's thesis, National Hsinchu University of Education). Retrieved from https://hdl.handle.net/11296/288qsf 16. Jhih-Hua, Chen, Yong-Song, Lai, Syun-Jhen, Jhang, & Zih-WEI, Yang.(2016) Feature Extraction and Machine Learning Methods for Liver Disease Detection and Prediction. Journal of Gerontechnology and Service Management, 4(3), 417-430. 17. Chien-Nan, Chen, & Yi-Lung, Yeh.(2010) Recovery Evaluation of Hillslope Disaster in Mountain Tribes. Journal of Taiwan Agricultural Engineering, 56(1), 61-70. 18. Szu-Yin, Chen(2016). QPESUMS Real-time Rainfall Forecasting Using Machine Learning Techniques (master's thesis, National Cheng Kung University, Tainan). Retrieved from https://hdl.handle.net/11296/29k9q9 19. Yu-Jhih, Chen, Guo-shu, Fan, & Jhao-Lang, Su.(2009) Research of Disaster Risk Assessment. Journal of Crisis Management, 6(1), 41-50. 20. Jun-you, Chen(2012). Wufeng Township, Hsinchu County refuge in evacuation planning of mechanism and asylum (master's thesis, National Central University, Taoyuan). Retrieved from http://etd.lib.nctu.edu.tw/cgi-bin/gs32/ncugsweb.cgi/login?o=dncucdr&s=id=%22NCU9833 02024%22.&searchmode=basic 21. Yu-Hua, Chen(2017). An Evaluation of the Isolation Effects in Mountain Community (master's thesis, National Taipei University of Education). Retrieved from https://hdl.handle.net/11296/x7gxb6 22. Bo-Jhen, Huang(2015). Radar Precipitation Estimation Using Support Vector Machine (master's thesis, University of Taipei). Retrieved from https://hdl.handle.net/11296/8c6szu 23. Wen-hsien, Yang (2010). Evaluation Processes of Bridge Hazards and Inventory of Bridges Damaged by Disasters (master's thesis, National Central University, Taoyuan). Retrieved from https://hdl.handle.net/11296/v48xxh 24. Jhen-Yu, Chen, Chun-Ming, Huang, Wun-Jhou, Huang, Jin-Tong, Jheng, Zun-Ying, Shih, & Shiu-Rong, Yang.(2010, October) Assessment of Isolated Community effect on settlements in Loaning River after typhoon Morakot. Taiwan Rock Engineering Symposium 2010, Kaohsiung. 25. Yun-Xing, Pan (2016). Discussion of the Rural Settlements Isolation Effects Occurs (master's thesis, National Taipei University of Education). Retrieved from https://hdl.handle.net/11296/jwa9bj 478

26. Geng-Han, Lin, Mei-Ting, Liu, Yu-Lun, Jheng, Yun-Xing, Pan, & Yuan-Fang, Tsai.(2017, October) Risk Clustering of Isolated Community Disaster Potential and Making Warning Threshold Of Rainfall. The 21th Cross-strait Hydraulic and Technology Seminar, Jiangxi. 27. Yu-Ciang, Fang, Jyun-Liang, Hsing, Pei-Cin, Li, Tong, Liu, Li-Ning, hang, Jian-Bin, Sie, ... Jyun, Hu (2016). The Deep Reading of the most important 20 Visual Machine Learning Theories for changing the next 20 years. Taipei, Taiwan: Top Team Information. 479

ISSSM-0262 Effect of Cognitive-Motor Training on Eye-Hand Coordination and Cognitive Function in Older Adults Kuei-Ru Chou School of Nursing, College of Nursing, Taipei Medical University, Taipei, Taiwan E-mail: [email protected] 1. Background Poor eye–hand coordination is associated with the symptoms of the early stage of cognitive decline. However, previous research on the eye–hand coordination of older adults without cognitive impairment is scant. Therefore, this study examined the effects of cognitive-motor training on the visual-motor integration, visual perception, and motor coordination sub-abilities of the eye–hand coordination and cognitive function in older adults. 2. Methods A double-blind randomized controlled trial was conducted with older adults. 62 older adults were randomly assigned to the experimental (cognitive-motor training) or active control (passive information activity) group, and both groups received 30 minutes of training each week, three times a week for 8 weeks. The primary outcome was eye–hand coordination, which was further divided into the sub-abilities of visual–motor integration, visual perception, and motor coordination. The secondary outcome was cognitive function. The generalized estimating equation was used to examine differences in immediate posttest, 3-month posttest, and 6-month posttest results between the two groups. Additionally, the baseline effect sizes were compared with the effect sizes of the immediate posttest, 3-month posttest, and 6-month posttests for the experimental group. 3. Results The visual–motor integration results showed a small to moderate effect size. The effect size of visual perception was small, whereas that of motor coordination was moderate to large. However, there was no statistically significant differences between the two groups. In addition, the results revealed statistically significant differences and small effect size in immediate posttest results for the attention dimension of cognitive function (p = 0.04). Keywords: Cognitive-Motor Training; Eye-Hand Coordination; Older Adults; Cognitive Function; Randomized Control Trial. 480

481