Application Prediction Model Using Identity Information
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing application prediction models struggle to accurately predict and preload newly installed applications, leading to slower startup times due to insufficient training data and lack of identity information for newly installed apps.
Innovation Solution
A method and apparatus for establishing an application prediction model that determines newly installed apps within a time window and trains a machine learning model using preorder use sequences and identity information, enabling improved prediction and preloading of target applications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional application prediction models are used without identity information, then the model structure is simpler, but the prediction accuracy for newly installed applications is insufficient
Solution Approach 1:
The patent applies preliminary action by determining identity information (whether an application is newly installed) before training the prediction model. The system collects installation time data and preorder use sequence data in advance, then uses this pre-prepared information to train the model, improving prediction accuracy for newly installed applications without requiring complex real-time analysis
Solution Approach 2:
The patent segments the training data by creating distinct identity information categories (newly installed vs. previously installed applications). This segmentation allows the model to learn different usage patterns for different application types, improving overall prediction accuracy while keeping the model structure manageable through organized feature categories
2Measurement precision
If more training data is collected to improve model accuracy, then prediction precision improves, but the data processing time and computational resources increase
Solution Approach 1:
The patent extracts only the most critical features needed for prediction: identity information (newly installed status) and preorder use sequences. By taking out and focusing on these key elements rather than processing all possible application data, the system achieves high prediction precision while minimizing data processing time and computational overhead
Solution Approach 2:
The patent applies partial action by collecting and processing only the essential subset of data required for accurate prediction. Instead of analyzing complete user behavior patterns across all applications, the system focuses on specific preorder sequences and installation status, achieving sufficient accuracy with reduced processing requirements
3Speed
If the system preloads applications based on accurate prediction, then startup speed improves, but system resource consumption increases
Solution Approach 1:
The system performs preliminary action by preloading applications only when the prediction model identifies high-probability candidates. By using identity information and preorder sequences to predict which applications will be launched next, the system preloads resources in advance for specific predicted applications, achieving fast startup speed while avoiding unnecessary resource consumption from blanket preloading of all applications
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A method of establishing an application prediction model comprising: determining (101, 501) a first application running in foreground at a sampling time in a preset sampling period; determining (102) whether the first application is installed within a time window having a preset time length and ending with the sampling time, to obtain identity information of the first application; and training (103, 504) a preset machine learning model based on sample data corresponding to the first application and a sample identity of the sample data, wherein the sample identity of the sample data includes the identity information of the first application and the first application.