Recommendation Engine Using Aggregate Distance Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network services struggle to provide personalized recommendations that accurately reflect a user's preferences based on their history of positive and negative interactions, often relying on popularity rather than individual tastes, leading to a lack of heterogeneity in recommended content.
Innovation Solution
A method that determines an aggregate positive distance and an aggregate negative distance for candidate objects by calculating commonalities with objects the user has positively and negatively interacted with, using outlier values to account for the breadth of interactions, thereby providing personalized recommendations based on user-specific preferences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If popularity-based recommendations are used, then recommendation simplicity is improved, but recommendation personalization accuracy deteriorates
Solution Approach 1:
The system changes the parameter for recommendation generation from popularity-based metrics to preference-based metrics. By calculating preference values based on user interactions with content items and comparing them against threshold values, the system transitions from simple popularity ranking to personalized preference matching, thereby improving personalization accuracy while maintaining operational simplicity through automated threshold-based decision making.
2Measurement precision
If recommendations are based on user interaction history, then personalization accuracy is improved, but system complexity deteriorates
Solution Approach 1:
The system extracts and isolates the essential components needed for personalization: user interaction data, preference value calculation logic, and threshold comparison mechanisms. By separating these core functions from the rest of the system, the patent reduces overall system complexity while maintaining high personalization accuracy through focused processing of only the necessary interaction history elements.
Solution Approach 2:
The system implements self-service by automatically generating personalized recommendations based on user interaction history without requiring manual curation or complex external processing. The automated preference value calculation and threshold comparison enable the system to serve itself in generating personalized content recommendations, reducing the need for complex manual intervention or external systems.
3Adaptability or versatility
If outlier values are used to account for interaction breadth, then recommendation heterogeneity is improved, but calculation complexity deteriorates
Solution Approach 1:
The system introduces outlier value parameters to the preference calculation process. By incorporating these outlier values as additional parameters in the preference value calculation, the system enhances its ability to account for the breadth of user interactions and generate more heterogeneous recommendations. This parameter enhancement improves adaptability while the structured nature of outlier value integration keeps calculation complexity manageable through defined computational steps.
Data Source
AI summary
Exemplary methods, apparatuses, and systems receive a candidate object with which a user can interact within a network service. For each of a first plurality of objects with which the user has had a positive interaction, a first value representing a commonality between the candidate object and each of the first plurality of objects is determined. For each of a second plurality of objects with which a user has had a negative interaction, a second value representing a commonality between the candidate object and each of the second plurality of objects is determined. An aggregate positive distance is determined using a plurality of the first values. An aggregate negative distance is determined using a plurality of the second values. The candidate object is displayed or not displayed to the user as a recommendation based upon a difference between the aggregate positive distance and the aggregate negative distance.