Method for recommending application based on user habit, mobile terminal, storage medium

CN118332192BActive Publication Date: 2026-05-29SHANGHAI DROI TECH CO LTD
2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
SHANGHAI DROI TECH CO LTD
Filing Date
2024-05-07
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

In a cloud phone environment, users find it difficult to quickly find apps that meet their needs. Existing technology leads to time-consuming and suitable apps being hidden and difficult to discover.

Method used

By collecting user-specific attribute information, application interaction information, and location information, analyzing and quantifying them, and then labeling them, user interaction scenario information is established. Then, neural networks and SVM algorithms are used to recommend applications that match the sentiment classification tendency. By combining sentiment analysis and interaction scenario matching, the time cost for users to find applications is reduced.

Benefits of technology

It enables the recommendation of suitable apps based on user habits, reducing the time cost for users to find apps and improving the accuracy and efficiency of recommendations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118332192B_ABST
    Figure CN118332192B_ABST
Patent Text Reader

Abstract

The application provides a method for recommending applications based on user habits, a mobile terminal and a storage medium, wherein the method steps comprise: collecting user inherent attribute information, application interaction information and position information, performing analysis and quantization, marking, and establishing user interaction scene information; collecting input information in an application interaction process, recording as interaction correlation data according to corresponding time periods and application package names; processing the interaction correlation data through a first neural network to output emotion classification; processing the interaction scene information through a second neural network to output real label probability distribution information; matching each reference interaction scene in a corpus according to the real label probability distribution information to find corresponding application programs; and performing emotion classification on the reference interaction scenes of each application program based on an SVM algorithm to recommend application programs meeting the emotion classification tendency to the user. In this way, the user can help to screen out APPs that meet the user application interaction habits, and the time cost of the user to find the application is reduced.
Need to check novelty before this filing date? Find Prior Art