A method to alleviate data sparsity in recommendation system based on step-by-step dynamic filling
A recommendation system and data sparse technology, applied in data processing applications, complex mathematical operations, instruments, etc., can solve problems such as sparse scoring data matrix, inaccurate merchant recommendation lists, etc., achieve accurate similarity calculation, increase the number of common scores, Fill full effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0047] On the basis of the traditional collaborative filtering recommendation algorithm based on users and items, the embodiment of the present invention adds the consideration of the average value of the historical common rating difference of users and businesses to the selection of similar neighbor sets, and finally dynamically and step-by-step. The unrated data in the business rating matrix is filled, and the method includes the following steps:
[0048] 101: Preprocess the user behavior data, and establish a user-food business rating matrix;
[0049]102: Construct a collection of historical scoring records for each user and merchant; construct a user collection, and sort the users in the user collection according to the number of merchants rated by users from large to small;
[0050] 103: Set user similarity threshold α and user history common score difference mean threshold β;
[0051] 104: According to the order of users in the user set, take a target user, calculate ...
Embodiment 2
[0062] The scheme in embodiment 1 will be further introduced below in conjunction with specific calculation formulas and experimental data, see the following description for details:
[0063] 201: For the user behavior data obtained from the American review website yelp, only select the relevant information of food merchants and their users, such as figure 1 , figure 2 shown;
[0064] Create a user-food business rating matrix, such as image 3 shown.
[0065] 202: Construct a user's historical scoring record set I(u) for each user, and construct a merchant's historical scoring record set U(i) for each merchant;
[0066] Build a user set U, count the number of merchants rated by each user, and sort the users in the user set according to the number of merchants rated by users from large to small;
[0067] 203: Set user similarity threshold α and user common score difference mean threshold β;
[0068] 204: Select the target item, calculate the similarity of user ratings, an...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com