Sparse Weight Matrix Generation for Recommendation Learning Time Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing Sparse Linear Method (SLIM) for shortening learning time in recommending items is hindered by the time-consuming calculation of degrees of similarity between items, which prevents further reduction in learning time for generating prediction models.
Innovation Solution
An information processing method that generates a weight matrix by extracting elements from each column of a target matrix, calculating similarity between items based on these elements, and setting the weight matrix as a sparse matrix with nonzero elements corresponding to item intersections, thereby reducing the computational load for similarity calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If the Sparse Linear Method (SLIM) is used to generate a weight matrix for item recommendation, then the learning time is shortened compared to traditional methods, but the calculation of degrees of similarity between items still consumes significant time
Solution Approach 1:
The patent segments the target matrix by extracting only some elements from each column instead of using all elements. This segmentation approach reduces the computational load for calculating degrees of similarity between items, thereby shortening the learning time while maintaining acceptable recommendation accuracy
Solution Approach 2:
The patent applies partial action by calculating similarity degrees based on only some extracted elements from each column of the target matrix rather than all elements. This partial computation significantly reduces the time required for similarity calculations and weight matrix generation, achieving faster processing speed with only slight reduction in recommendation accuracy
2Measurement precision
If all elements of each column in the target matrix are used to calculate the degree of similarity between items, then the accuracy of similarity determination is improved, but the computational load and learning time increase
Solution Approach 1:
The patent extracts only some elements from each column of the target matrix that are necessary for calculating the degree of similarity between items. This extraction process removes unnecessary computational overhead while retaining the essential information needed for accurate similarity determination, thus reducing learning time without significantly compromising accuracy
Solution Approach 2:
The patent changes the parameter of element selection by extracting only some elements from each column instead of using all elements. This parameter change reduces the computational complexity of similarity calculations, leading to faster processing speed while maintaining acceptable measurement precision for item similarity
Data Source
AI summary
An information processing system according to the present invention is an information processing system that sets a weight matrix. The weight matrix is generated by learning using a target matrix that is a matrix including an action status on an item in each of a plurality of setting statuses as an element of a column, includes a weight corresponding to an intersection of items as an element, and is multiplied by the target matrix. The information processing system includes: a similarity degree calculating unit configured to extract, from each column of the target matrix, some elements from among all elements of the column, and calculate a degree of similarity between the items based on the some elements of the each column; and a weight matrix setting unit configured to set the weight matrix that is a sparse matrix including a nonzero element based on the degree of similarity.