Set Based Item Recommendation System Using Precomputed Similarity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing recommendation systems are either fast but lack sophistication, or complex and time-consuming, often returning outdated results and requiring constant human input for rule maintenance, and they typically provide recommendations based on a single item rather than a set of items.
Innovation Solution
A system that receives a target set of items and queries a database to retrieve similar item sets, generating recommendations by performing a union of these sets, allowing for real-time, personalized recommendations based on item sets rather than individual items, and weighting recommendations by similarity and relevance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If recommendation systems use complex algorithms to provide sophisticated recommendations, then recommendation quality is improved, but processing time increases and results become outdated
Solution Approach 1:
The system pre-computes and stores item set similarities in a database before they are needed for recommendations. When a user requests recommendations, the system simply retrieves pre-computed similar item sets from the database and performs a union operation, avoiding time-consuming real-time complex calculations while maintaining sophisticated recommendation quality
Solution Approach 2:
The system merges multiple similar item sets together using a union operation to generate comprehensive recommendations. Instead of relying on a single complex algorithm, it combines results from multiple pre-computed item sets that share common items with the target set, achieving both speed and sophistication
2Speed
If recommendation systems provide recommendations based on a single item, then processing is fast, but recommendation accuracy and personalization are limited
Solution Approach 1:
The system transitions from single-item recommendations to item-set recommendations by adding a new dimension of analysis. Instead of finding similar items for one target item, it finds similar item sets that share multiple common items with the target set, providing more accurate and personalized recommendations while maintaining speed through pre-computation
Solution Approach 2:
The system pre-computes and stores multiple item sets with their similarities in a database before they are needed. When generating recommendations, it retrieves these pre-computed sets and performs a union operation, avoiding time-consuming real-time calculations while providing comprehensive multi-item recommendations
3Measurement precision
If recommendation systems use complex rules and algorithms, then recommendation sophistication is improved, but system complexity and maintenance requirements increase
Solution Approach 1:
The system automatically computes and stores item set similarities in the database without requiring constant human intervention. The similarity computation is performed once during data preparation, and the system then uses these pre-computed values for all recommendation queries, eliminating the need for ongoing complex rule maintenance
Solution Approach 2:
The system combines multiple simple operations (retrieving pre-computed similar item sets and performing a union) to achieve sophisticated recommendations. This approach avoids the need for complex real-time algorithms while maintaining recommendation quality through the aggregation of multiple item set results
Data Source
AI summary
A system for generating one or more recommended items for a user receives a target items set for the user which includes two or more target items. The system then queries a database of a plurality of related items sets using the target items set, where each related items set comprises a plurality of related items. In response to the querying, the system retrieves one or more similar items sets from the plurality of related item sets, each similar items set being similar to the target items set. The system then performs a union of the similar items sets to generate the recommended items.