This doctoral dissertation provides developed model for defining the patterns of users’ behavior in data traffic offloading while using smartphones and related applications. Provided model includes precisely defined phases and steps of research work related to the specific procedures of collecting, gathering, anonymization, processing and analysis of given data. The research is based on the data analysis of individual smartphone users’ characteristics and on the amount of offloaded data traffic from mobile networks onto Wi-Fi networks assessed by integrated application.

The individual characteristics of the smartphone users are determined based on answers of the on-line survey. To determine the proper and precise users’ characteristics relevant factors have been identified, affecting the generation of the smartphone data traffic and users’ preferences for transferring data traffic while using smartphone devices and related applications. Based on the identified factors and determined users’ preferences, the questions for the survey have been specified directly aiming user’s characteristics having a potential to become significant parameters in the development of model, as independent variables of the research.

The measurement of the generated data traffic has been conducted through the integrated smartphone application based on Android operating system. The real data of generated data traffic from the users’ smartphones include information about the amount of offloaded data traffic from mobile onto W-Fi networks. The users have been segmented based on the amount of the Wi-Fi offload into five categories, making a dependent variable of the research. The anonymization and merging of the collected information about generated data traffic of smart mobile devices and the users’ characteristics is being made. The research has included the analysis of 298 regularly and entirely fulfilled records of joint data, and the participants of the research have been selected by the method of appropriate sampling.

The fully developed model has proven the mathematical method of ordinal logistic regression as a useful method for the analysis of interdependence of joint data on the characteristics of users gathered by the survey providing the information on the amount of offloaded smartphone data traffic, segmented into five categories of users. Based on the phases and steps of the model development, the users’ preferences have been determined, which, in accordance to the values of the parameters of the model, represent the independent variables with a significant impact on the dependent variable, i.e. users category, determined by the amount of data traffic being offloaded from mobile onto Wi-Fi networks.

Provided results of the research and the given model affirm the probability that the users of certain characteristics (marked as significant) will belong to the one of the five determined groups of users, classified by the amount of offloaded mobile data traffic from mobile to Wi-Fi networks while using smartphones, thus defining the patterns of users’ behavior.

Back