Research And Application Of Deep Learning In Recommender System

Nowadays,with the advent of the era of big data,the amount of information is showing an exponential explosive growth trend.At the same time,the personalized needs of users are also increasing.In the face of massive data,how to help users get the information they need quickly and accurately is the main challenges currently facing.Therefore,the recommendation system has received more and more attention,and it has gradually become a core technology research module for data service companies and the entire internet industry.At present,the recommendation system has been widely used and has achieved certain results,but there are still some problems and deficiencies in the application,such as cold start caused by too little system data,data sparseness and other issues need to be researched and solved urgently,so it is of great significance to better research and develop recommendation system technology.In recent years,deep learning methods based on artificial neural networks have made major breakthroughs.Compared with traditional machine learning methods,deep learning has been applied in computer image and video,natural language processing and other fields and achieved better results.Therefore,people try to apply deep learning methods to the field of recommendation systems to carry out related research,and have achieved some results.Compared with the traditional recommendation algorithm,the deep learning method has stronger performance ability in the extraction of target features,it has better anti-noise ability to noise data,the deep learning method can learn the target features more accurately through neural network modeling,and can achieve better results.It also has a wider range of applicable scenarios.Combining with the existing research foundation and development direction,this paper studies the recommendation algorithm and deep learning method respectively,and introduces the deep learning technology into the traditional recommendation method,and then applies it in different recommendation fields to verify its reasonableness and effectiveness.First of all,this paper studies the development history of recommendation systems at home and abroad in recent decades and the application of deep learning technology in recommendation systems.This paper studies the basic algorithms commonly used in recommendation systems and the foundation of deep neural networks.Secondly,this paper conducts experiments on the traditional recommendation algorithm that is more popular at this stage,and introduces a variety of evaluation indicators to evaluate the results.The results reflect that the traditional recommendation algorithm has the problems of low recommendation accuracy,limited representation of features,and single application scenario.Combining the advantages of deep learning in feature extraction and large-scale data analysis,this paper proposes a feature aggregation recommendation model based on text convolutional neural network,and then configures and optimizes the model according to the characteristics of the data set and the training effect.Finally,this paper uses different data sets to apply the feature aggregation recommendation model to a variety of recommendation scenarios.The experimental results are displayed and comprehensively analyzed,which verifies that the feature aggregation recommendation model proposed in this paper has significant advantages over traditional recommendation algorithms in terms of accuracy and feature utilization.