Design And Implementation Of A Online Shopping System For Pet Supplies Based On Collaborative Filtering

Pets have always played an important role in people’s lives.With the development of social economy,the number of pet owners has increased significantly,and the consumption of pet supplies has also increased.Online pet supplies shopping platforms are more popular due to the mix of offline pet supplies stores,opaque prices and limited business hours.The rapid development of Internet technology has made online shopping more convenient and fast,but with the continuous growth of the quantity and variety of commodities,various commodity information floods the user’s field of vision,making it difficult for users to choose and wasting users’ time.In order to allow users to quickly obtain the goods they need from a large amount of data,an online shopping system based on personalized recommendation emerges as the times require and becomes an effective solution.This paper designs and implements an online shopping system for pet supplies based on collaborative filtering recommendation algorithm.The collaborative filtering recommendation algorithm calculates the similarity between users based on the interaction between users and items,and predicts other items that users may be interested in based on this.In some current studies,graph convolutional networks are introduced into models to measure the similarity between users,and good results are obtained.However,feature transformations and nonlinear activations in graph convolutional networks contribute little to the performance of collaborative filtering.Also,including them in the model increases the difficulty of training and reduces the effect of recommendation.Therefore,this paper introduces the Light GCN(light graph convolution network)graph convolution network model to implement collaborative filtering recommendation.By performing light graph convolution and layer group operations on the user-item interaction matrix,the feature representation of users and items is generated and internalized.Product operation,get the user’s predicted score for the item,generate the final recommendation list,and train on the public dataset Movielens.On the Movielens dataset,the model achieves a recall rate of 61.26% and an NDCG of 59.42%,which is in line with expectations.The system is mainly divided into a front-end subsystem and a back-end subsystem.The front-end subsystem includes four modules,a personal center module,an online mall module,a community interaction module,and a pet-raising knowledge module.The back-end subsystem includes a back-end management module.The front-end of the software uses JSP,Java Script,and CSS technologies,and the back-end uses Spring MVC,Spring,and My Batis technologies,that is,the SSM framework.At present,the system has completed the design and development,and completed the relevant testing work.