Application Quality Prediction via User Behavior Regression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for evaluating mobile application quality rely heavily on user reviews and download metrics, which are often inaccurate due to lack of user evaluations, and fail to fully utilize behavioral data for more precise assessments.
Innovation Solution
A regression algorithm is employed to predict application quality by extracting features from user management behaviors such as downloading, uninstalling, and updating, using data preprocessing, behavioral feature extraction, and model training to determine application quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If user evaluation (scoring and reviews) is used to determine application quality, then evaluation accuracy is improved, but the method becomes inapplicable when users are not accustomed to evaluating applications resulting in few user reviews
Solution Approach 1:
The patent introduces behavioral features (download, uninstall, update actions and their sequences) as an intermediary to indirectly measure application quality when direct user evaluations are unavailable. These behavioral patterns serve as mediators between user interaction and quality assessment, enabling evaluation without requiring explicit user reviews.
Solution Approach 2:
The patent replaces the mechanical system of explicit user scoring and reviewing with an automated analysis system that infers quality from implicit behavioral data. Instead of relying on users to manually evaluate applications, the system automatically processes download, uninstall, and update sequences to determine quality metrics.
2Difficulty of detecting and measuring
If simple indicators (download numbers, downloading user numbers) are used to measure application quality, then ease of measurement is improved, but the measurement precision deteriorates because these indicators may not truly reflect application quality
Solution Approach 1:
The patent segments the simple download metric into multiple behavioral components: download actions, uninstall actions, update actions, and their chronological sequences. This segmentation transforms a single coarse indicator into multiple fine-grained behavioral features that collectively provide more precise quality measurement while remaining computationally straightforward.
Solution Approach 2:
The patent adds temporal and sequential dimensions to the measurement by analyzing the order and timing of user actions (download → use → update → uninstall). This transforms a static count metric into a dynamic sequence-based measurement, capturing quality insights from behavioral patterns across multiple dimensions without significantly increasing measurement complexity.
3Measurement precision
If user management behavior data is fully utilized to determine application quality, then measurement precision is improved, but device complexity increases due to the need for data preprocessing, feature extraction, and model training
Solution Approach 1:
The patent performs preliminary actions by pre-processing user behavior data and extracting behavioral features (download, uninstall, update sequences) before quality determination. This preliminary processing organizes raw data into meaningful patterns, simplifying the subsequent quality assessment process and making the system more efficient despite the added initial processing step.
Solution Approach 2:
The system uses user-generated behavioral data to automatically evaluate application quality without requiring external intervention or complex manual analysis. The user's own download, uninstall, and update actions serve as the evaluation criteria, making the system self-sufficient and reducing the need for additional complex evaluation mechanisms.
Data Source
AI summary
The present invention relates to a method and system for determining the quality of an application based on user behaviors of application management. The method includes: 1) preprocessing data on user behaviors of application management to obtain a sequence of management behaviors as a training set; 2) extracting features of the user behaviors from the obtained sequence of management behaviors; 3) establishing a multi-dimensional independent variable space using the extracted behavior features, wherein each application in the training set is considered as a point in the space; training a model according to the space; and debugging to obtain optimal model parameters and a definite regression model; 4) arranging the data of the management behaviors of each application to be predicted according to step 1) and step 2), and then inputting the result into the regression model, wherein an output of the regression model is the result of the application quality determination. The present invention can effectively remove irrelevant features and exert the quality expression ability of relevant features, and can obtain more accurate results for evaluation of application quality.