App startup acceleration method based on heterogeneous network embedding with attributes

Abstract

The invention discloses an APP start-up acceleration method based on heterogeneous network embedding with attributes, aiming at predicting APP start-up and realizing start-up acceleration according to the predicted application. The technical solution is to build an application start-up acceleration system based on a heterogeneous network with attributes. The system consists of a data processing module running on a mobile terminal and a model training system running on the cloud; the data processing module consists of a data collection module, a data pre-set Processing module, data upload module, application startup prediction module, application startup acceleration module, terminal database; model training system consists of application record storage module, model training module, model interaction module and cloud database. The application startup acceleration system uses the APP startup prediction method embedded in the heterogeneous network with attributes to accurately predict the next possible startup application, realizes the creation of the application process in advance and loads the application resources in advance, and reduces the user's waiting time for the application startup.

Description

technical field [0001] The present invention relates to the field of APP (ie, application) start-up acceleration in an Android (Android) system, and in particular to a method for predicting APP start-up based on a heterogeneous network embedding method with attributes to accelerate APP start-up. Background technique [0002] In the Android system, a general application is composed of four major components: Activity (activity), Service (service), ContentProvider (content provider), and Broadcast (broadcast). In the Android system, it is responsible for the startup and switching of the four major components. , Scheduling and application process management and scheduling is ActivityManagerService (AMS, activity management service process). AMS is one of the core services in the Android system. Its responsibilities are similar to the process management and scheduling modules in the operating system, and the application startup process is closely related to it. figure 1 Start th...

AI Technical Summary

Problems solved by technology

This method can be applied to application prediction, but this method can only predict whether there will be a link relationship between two nodes, and can only make predictions based on a certain context such as time, location information, etc., resulting in the accuracy of the prediction method Not high, and previous studies have shown that time, location, and other contexts have an impact on the accuracy of application predictions

There is no report on this method of accelerating application launch through prediction

Images