A Study On The Influence Factors Of User Privacy Concern And The Principle Of Privacy Information Diffusion In Mobile Internet

The mobile Internet is the world’s fifth new technology cycle in the past half century. It allows users to access the Internet and use mobile applications from handheld mobile terminals, such as smartphones or tablet computers, via a mobile network or other wireless network. Compared with the traditional Internet, the mobile Internet has three remarkable characteristics:mobility, local and social. While the mobile Internet provides high-quality services, much of users’private information will be exposed. This information includes transaction information, the type of mobile services used, and the call history. Unavoidably, such private information may be used illegally or without acceptance of the owner. At the same time, the social network sites can predict the users’behavior through the method of data mining. All of these factors increase users’concerns about their privacy. Thus, users with different levels of concern about their privacy behave differently. As we know, it is important to do researches deeply in the area of privacy.Taken mobile Internet as the study object, the paper concentrates on the variables that influence users’privacy concern and the effects of personal characteristics and interpersonal influence on privacy information diffusion. There are three specific research contents:(i) Study on factors which influence users’privacy concern in mobile Internet.Compared with the traditional Internet, the mobile Internet has three remarkable characteristics:it is mobile, local and social. Based on the theory of TRA and CFIP, this study aims to investigate the factors that affect privacy concerns about the mobile Internet in China and which dimensions of privacy concerns these factors affect. Specifically, we propose a research model with six factors-including the institutional factor, personal disposition toward privacy, internal locus control, personal innovativeness, online activity and social influence-and investigate the relationship between the influence factors and privacy concerns. This part of the study used questionnaires to collect the corresponding user data, and estimated the path coefficient using the tools of SEM. The findings indicate that the six factors all have significant effects on all or some of the dimensions of privacy concerns and support the negative relationship between privacy concerns and the willingness to provide personal information.(ii) Research on the time of privacy information diffusion under the environment of mobile Internet.In mobile Internet, users with different privacy concern levels exhibit different private information posting behavior. We put forward a series of influence factors, including privacy information features, user features, and social network influence, and these effect are differ from the different levels of privacy concern of users. To investigate and contrast the effects of personal and interpersonal influence on private information diffusion for different levels of privacy concern of users, we establish a survival analysis model and examine influence factors, including privacy information features, user features, and social network features. The findings indicate that these three types of factors are very important and have a significant impact on the diffusion time of private information, especially the social network features. The influence factors differ for users with different levels of privacy concern. The results aid in our understanding and prediction of users’performance in mobile Internet, with implications for social marketing and public opinion monitoring,(iii) Predicting the scale of information diffusion in mobile InternetIn this part, we predict the scale of information diffusion in mobile Internet, and use the same influence variables like above. There are three categories:privacy information attributes, user attributes and social network attributes. Three classic data mining models are established, which are neural network, decision tree and linear regression. At last, we select an appropriate model that can fit real data better and have a higher accuracy. It is found that the RBF neural network model performs much better than decision tree and linear regression do. And the results can be helpful in our prediction of users’performance and predicting the scale of information diffusion in mobile Internet better.There are four innovation points in this research.(1) This study establishes models based on the mobile Internet environment. All of the influence factors and four dimensions of privacy concern (CFIP) are combined with the characteristics of the mobile Internet, which make the conclusion be more targeted and reflect the real situation of the mobile Internet environment. At the same time, the social influence is joined the influence factors, which is verified with empirical data and mature statistical analysis method.(2) This article study on four dimension of CFIP (collection, error, improper access and second use) separately. We analyze the influence of the six factors on the four dimensions, to make the enterprises have a better understanding on users’thoughts and habits. They can use corresponding privacy protection scheme and marketing program for better services.(3) Previous studies had not combined privacy and user information diffusion principle. This study fills this gap by studying social groups’ influence on users’privacy concern and behavior in the perspective of social network. Survival model is used and privacy information diffusion times of users on different privacy concern levels are compared.(4) Influence factors on privacy information diffusion time and scale are summarized into three categories. One category is the features of text including Average Words, Average Hashtags, Average Photos and Average URLs. Another category is the features of users including Gender, Age, verified, Social Activeness and Social Popularity. The last category is the features of social network structure including Degree, Similarity and Structural cohesion. This improves and deepens the research on the influence factors of privacy information diffusion.