User Model Rating Matrix for App Behavior Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to accurately infer and utilize user behavior features from mobile device data, as they are often general and do not provide relevant information for targeted advertising or app recommendations.
Innovation Solution
A computer system that builds a user model based on a rating matrix using selected rating parameters from user data, employing matrix factorization and Spark technology to compute user and app representations, enabling efficient and robust user behavior analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If user behavior data from mobile devices is collected and analyzed, then user model accuracy for advertising and app recommendations is improved, but data processing complexity and computational requirements increase
Solution Approach 1:
The patent segments user behavior data into distinct categories (app installation data, usage data, rating data) and processes each segment separately through specialized modules. This segmentation allows the system to handle complex data processing by breaking it down into manageable, organized components that can be analyzed independently and then integrated to form a comprehensive user model.
Solution Approach 2:
The patent introduces a matrix factorization module as an intermediary between raw user behavior data and the final user model. This intermediary component transforms complex, high-dimensional user behavior matrices into simplified latent factor representations, effectively mediating the transition from complex input data to accurate user models while reducing computational complexity.
2Reliability
If detailed user behavior data is processed to create accurate user models, then app recommendation quality is improved, but processing time and computational resources increase
Solution Approach 1:
The patent performs preliminary matrix factorization on user behavior data to pre-compute latent factor representations before actual recommendation generation. By preprocessing the data and extracting key latent features in advance, the system reduces the computational burden during real-time recommendation operations, thereby improving processing speed without sacrificing recommendation quality.
Solution Approach 2:
The patent transforms the original high-dimensional user behavior parameters into a lower-dimensional latent factor space through matrix factorization. This parameter transformation maintains the essential information needed for accurate recommendations while significantly reducing the computational complexity and processing time required for analysis.
3Quantity of substance
If matrix factorization is used to compute user features, then dimensionality reduction is achieved, but computational complexity of the factorization process increases
Solution Approach 1:
The patent applies partial matrix factorization by computing only the essential latent factors needed for recommendation tasks rather than performing complete factorization of all user behavior data. This partial action approach achieves sufficient dimensionality reduction for practical recommendations while avoiding the excessive computational complexity of full factorization.
Data Source
AI summary
Systems and methods are provided for building a user model. The system includes a processor and a non-transitory storage medium accessible to the processor. The processor is configured to obtain user data from a database, where the user data include user behavior for a plurality of apps installed on one or more user terminals. The processor selects at least one rating parameters using the user data, where the at least one rating parameters indicates a rating of relevant app usage. The system builds the user model based on a rating matrix comprising the at least one rating parameters.