Sparse Data Matrix Product Recommendation Boosting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional product recommendation systems face challenges with sparse data matrices, where insufficient historical electronic activity leads to unreliable association values for new or less popular products, resulting in repetitive recommendations.
Innovation Solution
The system aggregates products with similar characteristics using machine-learning models to determine similarity values and boosts the historical electronic activity of target products with sparse data entries by copying activity from similar products, recalculating association values to generate more accurate recommendations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional product recommendation systems use historical electronic activity to generate association values, then recommendations can be made based on user behavior patterns, but new or less popular products with insufficient historical activity receive unreliable or no association values, leading to repetitive recommendations
Solution Approach 1:
The patent applies the copying principle by copying historical electronic activity data from established products to similar products with sparse data. The system identifies products with similar characteristics (e.g., same category, attributes) and copies their interaction history to populate association values for products that would otherwise have insufficient or no historical data, thereby enabling reliable recommendations for new and less popular products.
Solution Approach 2:
The patent uses an intermediary approach by introducing a similarity measurement mechanism as a mediator between products with sufficient historical activity and products with sparse data. This intermediary layer (similarity calculation) enables the transfer of useful information from established products to similar products, resolving the data sparsity problem without directly connecting the two product sets.
2Adaptability or versatility
If the data matrix is used to store historical electronic activity for all products, then comprehensive product coverage is achieved, but the matrix becomes sparse with many entries having insufficient or no historical activity
Solution Approach 1:
The system copies historical electronic activity from products with sufficient data to products with sparse or no data by identifying similar products and transferring their interaction patterns. This copying mechanism fills in missing information in the data matrix, enabling comprehensive product coverage while maintaining reliable association values through information transfer from established products.
Solution Approach 2:
The patent applies preliminary action by pre-calculating product similarities and preparing to copy historical activity before actual recommendation generation. The system proactively identifies products with sparse data and transfers appropriate historical information in advance, ensuring that when recommendations are made, sufficient association values are already available for all products in the matrix.
Data Source
AI summary
A recommendation service access a data matrix listing of products associated with product profiles, the data matrix having product entries that store sparse historical electronic activity. For a target product it is determined which other products should be used to boost the historical electronic activity of the target product based on a first subset of product profiles that share product characteristics with the target product. Similarity scores are computed between the product profile of the target product and the first subset of product profiles to identify a second subset of one or more products having a similarity score above a scoring threshold. The historical electronic activity of the target product is boosted using the historical electronic activity of the other products in the second subset. Association values are calculated between the target product and the other products in the second subset by based on the boosted activity.