The Machine Learning Solution based on Period and Deep Construction of Mobile Data for Predicting User Habit

Information analysis on user habit is becoming hotter as its value for many potential and profitable areas. However, how to excavate hard-core data and obtain the most effective information associated with research target, and how to obtain more precise prediction of user habit timely and fast on mobile terminals, are still big challenges. Most of the traditional methods use static data, which can no longer meet the requirements of the movable era, such as learning daily travel route and predicting the next most possible applications. Additionally, though algorithms based on Artificial Intelligence (AI) have boomed, many researches only based on direct data and pay insufficient attention to the deeper interrelation of data. This paper introduces two relevant AI models improving the phone memories; then, the main solution is recommended in detail. It mines some closely correlated parameters by specific mechanism named Linked Trigger (LT) and filtering policies for positioning, digging their underlying relations for better learning user habit by special construction ‘Directed Location Pair Application (DLPA)’. Based theseanalysis on the deep-seated connection between the collected data, the conditional probability referring Bayesian network is used to learn and predict the location habit and application habit.