Survey-Based Ranking Model Selection for New User Recommendations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current software platforms fail to effectively personalize item recommendations for users, especially new users, as they lack sufficient data processing capabilities to accurately recommend relevant items based on user preferences, leading to inefficient computing resources and irrelevant suggestions.
Innovation Solution
A computer-implemented method and apparatus that programmatically selects a user survey data object and ranking models based on user constraints and engagement data, generating a customized output ranked item data object set by processing survey engagement data and applying user-defined filters, thereby reducing computing power and improving recommendation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional software platforms process user engagement data for item recommendations, then recommendation coverage is provided, but recommendation accuracy deteriorates for new users due to insufficient data
Solution Approach 1:
The system performs preliminary actions by selecting and presenting survey questions to new users before actual purchasing behavior occurs. This proactive data collection approach allows the system to gather preference information in advance, enabling accurate recommendations even when historical engagement data is insufficient. The survey mechanism pre-emptively addresses the data scarcity problem for new users.
Solution Approach 2:
The patent introduces survey questions as an intermediary mechanism between the user and the recommendation system. These questions serve as a mediator that translates user preferences into structured data, bridging the gap between new users with no purchase history and the recommendation algorithm. The survey acts as a proxy for the missing engagement data.
2Adaptability or versatility
If software platforms collect and process extensive user engagement data, then recommendation personalization improves, but computing resource consumption increases
Solution Approach 1:
The system extracts only the most essential preference information needed for recommendations by selectively presenting targeted survey questions rather than collecting comprehensive engagement data. This extraction approach gathers sufficient personalization data while minimizing computing resource consumption associated with processing extensive user interaction histories.
Solution Approach 2:
The patent applies partial action by collecting only the necessary subset of user preference data through surveys, rather than processing all possible engagement data. This selective data collection achieves adequate personalization without the computational overhead of analyzing complete user interaction patterns, especially for new users.
3Loss of information
If survey questions are presented to users, then preference data collection improves, but user interaction time increases
Solution Approach 1:
The system implements partial action by presenting only a selective subset of survey questions to users rather than administering complete surveys. This approach collects sufficient preference information to generate accurate recommendations while significantly reducing the time burden on users. The system prioritizes questions based on their informational value for recommendation accuracy.
Solution Approach 2:
The patent segments the survey into discrete, manageable questions that can be answered quickly. By breaking down preference collection into small, focused units rather than presenting lengthy comprehensive surveys, the system minimizes user interaction time while still gathering essential data for personalization.
Data Source
AI summary
Embodiments of the present disclosure provide mechanisms for selection of a user survey data object from a set of user data objects, and processing of survey engagement data associated with a selected user survey data object. The user survey data object selected is appropriate for providing associated with a particular user data object, and the survey engagement data received associated therewith enables programmatic selection and use of particular ranking model(s) for use in generating and providing an output ranked item data object set. Example embodiments utilize selected ranking model(s) of a set of ranking models to programmatically generate and output an output ranked item data object set for a particular user profile.