An Item-Based Content Recommendation Method

Abstract

The invention discloses an item-based content recommendation method, which includes the following steps: 1) modifying the initial scoring value, and sorting the original matrix E into E m×n ; 2) Using the Pearson correlation formula, traversing and substituting E m×n Each element of , so as to obtain the corresponding Pearson item similarity matrix item‑SIM m×n ; 3) For the normalized scoring matrix E m×n Do matrix singular value decomposition to get U, S, D; 4) Determine the appropriate dimension k to get matrix U k , S k 、D k ; 5) Calculate the user-item scoring matrix E according to the formula m×n The scoring similarity matrix E k , as the entry matrix for subsequent calculation of predicted ratings; 6) Traversing each user rating row vector E that has rated item j u , if there is a value at item i, then substitute the following formula to calculate the average deviation value dev of the item j,i : 7) will get the average deviation value dev j,i , project similarity similarity i,j Substitute to get user u to item j The recommended prediction score of Prediction IS‑SVD (u) j ; The present invention can be adapted to data of different densities, and the adaptability of the algorithm is enhanced.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to an item-based content recommendation method. Background technique [0002] Today, with the development of Internet technology, more and more people buy and consume goods through the Internet, and among so many users and goods, accurately calculating which users are interested in which goods is also an urgent problem in the computer field. problem. In order to solve this problem, various personalized recommendation systems have been introduced in the computer field. [0003] In recent years, popular recommendation algorithms can be roughly divided into item content, collaborative filtering, user knowledge, and association rule recommendation directions. Among them, an item-based collaborative filtering (Item-based CF) slope one algorithm and a weighted slope one algorithm that adds the number of scoring users of the item can solve the above problems well. The algorithm predic...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art, and to provide an item-based content recommendation method, which can solve the problem of problems due to no internal connection between items while retaining as much scoring information as possible. The problem of poor accuracy of the recommendation system due to non-dense data sets

Images