Application Recommendation Sorting via Download and Usage Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current application recommendation methods often mismatch the popularity and prevalence of applications with their recommendation positions, leading to inappropriate recommendations, as they primarily rely on download times rather than user usage patterns.
Innovation Solution
A method that determines and combines download times and user use times, converts them to a comparable order of magnitude, performs weighting calculations to obtain sorting values, and uses these values to sequence applications on a recommendation interface, ensuring that popular and prevalent applications are prioritized.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If application recommendation position is set based on download times, then applications with high download times can be easily identified and positioned, but applications with make-up download times but low actual popularity will occupy front positions inappropriately
Solution Approach 1:
The patent changes the parameter used for recommendation positioning from download times to user use times. This parameter change allows the system to measure actual application popularity and prevalence more accurately, resolving the issue where make-up download times incorrectly positioned low-popularity applications at front recommendation positions.
2Measurement precision
If application recommendation position is set based on user use times, then the popularity and prevalence at user level can be accurately reflected, but the complexity of data collection and processing increases
Solution Approach 1:
The patent implements a self-service mechanism where terminal devices automatically report their application usage data to the server. This eliminates the need for complex manual data collection systems, as the data is gathered automatically from the source (user devices) through standardized reporting protocols.
Solution Approach 2:
The system establishes a feedback loop where terminal devices continuously report usage data to the server, which then updates recommendation positions based on this feedback. This automated feedback mechanism enables accurate tracking of user use times without requiring complex manual intervention or processing systems.
3Adaptability or versatility
If multiple metrics (download times and user use times) are combined for recommendation positioning, then the comprehensiveness of recommendation criteria is improved, but the complexity of calculation and processing increases
Solution Approach 1:
The patent applies partial action by selectively using user use times as the primary sorting criterion while treating download times as a supplementary or secondary factor. This approach maintains comprehensiveness by considering multiple metrics but reduces calculation complexity by establishing a clear hierarchy of importance rather than requiring equal processing of all metrics.
Data Source
AI summary
In some embodiments, an application recommendation method includes: determining download times and user use times of each of multiple to-be-recommended applications; converting the download times and the user use times to be in a same order of magnitude, to obtain converted download times and converted user use times; performing weighting calculation on the converted download times, to obtain a download sorting weighted value of the application, and performing weighting calculation on the converted user use times, to obtain a use sorting weighted value of the application; obtaining a sorting value of each of the applications based on the download sorting weighted value and the use sorting weighted value of the application; and determining a sequence of the multiple applications on an application recommendation interface based on the sorting value of each of the applications.