Current social networking services employ users' social graphs to recommend friends for users to connect with. It's possible that this doesn't reflect a user's true preferences for friends, though. Friend book is a game-changing approach to expanding your social circle. Friend Book does not utilize your social network to assist in friend discovery. Instead, it makes advantage of your regular routine. Friend Book analyzes user-centric sensor data to learn about users' habits and the degree to which they are like those of other users; it then makes recommendations to those who most closely resemble its model users. To determine how a user lives, we use the perspective that the user's everyday activities are life documents and apply a technique called Latent Dirichlet Allocation, which is inspired by text mining. In addition, we provide a friend-matching system and a similarity metric for analyzing the degree to which users' behaviors are similar. If a user asks a query and receives a lot of helpful suggestions, Friend Book will show them who those people are. Users can now provide feedback in Friend Book. Because of this, the app's recommendations improve. We tested Friend book's performance on Android-powered smartphones in both small and large-scale simulations to determine its viability. The data demonstrates that the recommendations accurately portray what people look for in a buddy.

Current social networking services employ users' social graphs to recommend friends for users to connect with.