Evaluating the Role of Location-Based Context in Predicting Mobile App Usage Behavior

Abstract

The behavior of mobile apps usage is also needed as it can assist a developer to raise the involvement of users to their mobile app and polish mobile experience. One of the key information that determines a user behavior is the location-based context and this may be used to indicate the user input made with applications depending on the place that he/she is in. In this case, the study answers the question concerning the predictive capacity of location-based context when it comes to mobile applications, or, in other words, how sites of information (such as those portrayed by the coordinates of the global positioning system, distance to the landmarks, and time) can assist in the enhancement of the app use predictions. Previous research studies have shown the significance of the contextual information on the prediction of the user behavior (Zhao et al., 2019; Huang et al., 2020). Random Forest and Support Vector Machines (SVM) are the machine learning algorithms we perform to analyze the application usage patterns on the basis of location data within our works. The predictions on basis of user behavior only instead of running model at location-based context have also been suggested as the quality of prediction accuracy will be much better. The study suggested is significant since the availability of literature on context-aware computing presents a lot, particularly on the effectiveness of the location-based features in predicting app use. The implications of such an approach are huge to the producers of the apps, since this can extend the tendency towards personal user experience and tailor-made advertising initiatives.