A recommendation model based on Stepped Six Channel CNN feature extraction

In a recommendation system, users and projects often do not exist independently. There are a lot of potential factors. Although the existing mainstream models can extract the potential features between users and projects, the extracted potential features are mainly high-order features and original f...

Full description

Saved in:
Bibliographic Details
Published in:Journal of physics. Conference series 2022-09, Vol.2347 (1), p.12009
Main Authors: Chen, Chuang, Zheng, Shanhong, Zhan, Qi, Guo, Jia, Chen, Jia
Format: Article
Language:English
Subjects:
Channels
Feature extraction
Physics
Recommender systems
Standard data
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In a recommendation system, users and projects often do not exist independently. There are a lot of potential factors. Although the existing mainstream models can extract the potential features between users and projects, the extracted potential features are mainly high-order features and original features. We believe that there are many features at the intermediate level and the interaction between many features. Therefore, we proposed a Stepped Six Channel CNN network model (SSCC). By using different network depths to process the interaction information of user information and project attributes, the potential characteristics of multiple layers can be extracted, and feature layers of different channels are combined to form new feature channels. In addition, the network model based on DNN is used to predict the interaction between the user and the project, and the two models are combined in parallel to form a new model which name is Stepped Six Channel CNN&DNN(SSCCD). Our proposed model is clearly superior to some of today’s mainstream models through experimental tests on publicly available standard data sets.
ISSN:1742-6588
1742-6596
DOI:10.1088/1742-6596/2347/1/012009