Recommending method integrating user and project rating and characteristic factors

Abstract

The invention relates to a recommending method integrating user and project rating and characteristic factors, which mainly aims at solving the problems of cold starting and data sparsity in traditional collaborative filtering recommending algorithm. A selected experimental data set is classified into a training data set and a test data set; similarity based on users and user characteristics and similarity based on projects and project characteristics are respectively calculated on the training data set; the similarities based on the users and the user characteristics are combined by selecting appropriate weight; a user nearest neighboring matrix is calculated respectively according to a dynamic selection threshold value method and a K nearest neighbor (KNN) method, an optimum method is selected through the comparison, and user recommend results are calculated by utilizing a recommending formula; the similarities based on the projects and the project characteristics are combined by selecting an appropriate weight; a project nearest neighboring matrix is calculated respectively according to the dynamic selection threshold value method and the KNN method,, an optimum method is selected through the comparison, and project recommend results are calculated by utilizing a recommending formula; and the user recommend results are combined with the project recommend results by selecting an appropriate weight.

Description

technical field [0001] The invention is a recommendation algorithm generated by integrating scoring data and characteristic factors of users and items, which mainly solves the cold start problem and data sparsity problem in traditional recommendation algorithms, and belongs to the technical field of recommendation research. Background technique [0002] Most of the techniques used in recommender systems today are mainly divided into two categories: content-based recommendation algorithms and collaborative filtering recommendation algorithms. The content-based recommendation algorithm refers to the calculation of similarity based on the attribute feature information of the item itself, while the collaborative filtering recommendation algorithm refers to finding users with similar interests to the specified user in the user group by analyzing user interests, and synthesizing the similarity of similar users to a certain information. The evaluation of the system forms the predic...

AI Technical Summary

Problems solved by technology

[0003] There are three common problems in existing recommendation algorithms: (1) Data sparsity problem: when the number of items is large but the number of items rated by users is small, a very sparse rating matrix will be generated

(2) Cold start problem: new users have not rated any items, and new items have no users who have rated them

(3) Scalability problem: When the system scale becomes larger, the performance of the recommendation algorithm drops sharply, resulting in inaccurate recommendations

[0004] Collaborative filtering algorithm is generally considered to be the most successful recommendation algorithm, but it cannot solve the common cold start problem in traditional recommendation algorithms. When there are new users and new items, it is difficult to generate them through collaborative filtering algorithm because there is no score record. Good recommendation effect, so a recommendation algorithm that can be generated based on user characteristics and item characteristics is needed

At the same time, in real applications, the number of items is huge, but the items rated or purchased by users are small, which leads to a very sparse user rating matrix, which directly leads to inaccurate calculation of similarity between users or between items, thus affecting the recommendation accuracy. Therefore, it is necessary to predict user ratings through recommendation algorithms to solve the problem of data sparsity

[0005] In the recommendation algorithm, too much attention is paid to the user or item characteristic factors and the rating data is ignored. Due to the limited item characteristics, it is difficult to obtain more data effectively, while too much attention is paid to the rating data and the user or item characteristic factors are ignored because it is based on historical data. , there will be a cold start problem for new users or new items, so biasing towards either side may lead to a decline in the performance of the recommendation algorithm, so it is necessary to consider the two comprehensively, and use weights to balance the influence of feature factors and scoring data in the recommendation algorithm degree

Images