Definitions of Smartphone Security Strategies by SVM Based Market Segmentation and Hybrid MCDM Methods

• Main Authors: Pai, Fu Nien, 白傅年
• Other Authors: Chang, Shih Yu
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/36699617565383673977

碩士 === 國立清華大學 === 通訊工程研究所 === 104 === Due to the mobile communication devices with convenience, positioning and mobility features, smartphones are becoming essential part of the communication system in human life. Therefore, the information security of smartphone grows with each passing day. Even though there are some scholars probed for smartphones information security problems. But researchers rare for using market segmentation to analysis the difference between customer behaviors, moreover, discussion about smartphone information security and its solution projects further. In order to work out this problem, this research first based on questionnaires to investigate smartphone usage behaviors, and according the support vector machine (SVM) method to cluster, which based on users mode as market segmentation. Through market segmentation, the principal component analysis (PCA) that induce and simplify each market segment’s customers characteristic. Through the way of the PCA method, using the rough set theory (RST) to conclude the rules with affected influencing smartphone usage behaviors. Next, this research takes the decision making trial and evaluation laboratory (DEMATEL) method to construct the causal relationship among the criteria. In view of the DEMATEL method, construct customer characteristics influence relationship map (IRM) diagram for each segment. The analytic network process (ANP) generates the relation diagram on account of the analysis framework by IRM. To infer the weight mapping to each criterion, collecting expert opinions and general questionnaires simultaneously. In the light of expert opinions and general questionnaires, this can produce security strategies of each market separation by client actions. Stand on the results of this research, this research will provide substantial suggestion consulted for smartphone antivirus softwares.