Research And System Implementation Of Collaborative Filtering Recommendation Algorithm For Home Smart Cloud Audio

With the rapid rise of high-tech industry,embedded terminal devices are widely used in the field of smart home.As an important representative of smart home,home audio system has become an important tool for people to relax.With the growing maturity of cloud computing,many music radio enterprises can provide more abundant resource support for products through cloud services.At the same time,personalized recommendation has become one of the necessary functions to improve product satisfaction.In the era of information overload,helping users quickly locate the information in line with their own interests has received a lot of research,but also has precious commercial value.Therefore,the development of a set of home intelligent cloud audio system with recommendation service meets the actual demand of the market.This thesis studies collaborative filtering recommendation algorithm based on improved user similarity.At the same time,based on the user reqirement analysis,the home intelligent cloud audio system is modularized.And the improved algorithm is applied to the system.In the research of algorithm improvement,the calculation of user similarity has to face the problems such as the difference of co-rated item number from common users,the difference of rating value and the difference of item popularity.To deal with these problems,a new method of calculating user similarity is proposed.Firstly,the difference of co-rated item number rated by common users could be relieved by the proposed method by combining the weighted cosine and modified cosine similarity.Secondly,the proposed method further exploits two correction factors to reduce the impact of numerical difference and item popularity difference on user similarity calculation,which reduce the error of rating prediction.Finally,the experiment conducted on the Movielens dataset shows that the proposed method outperforms the baseline methods in terms of Mean Absolute Error(MAE)while improving the accuracy of user rating prediction and recommendation quality with excellent robustness.This thesis will introduce the algorithm in one chapter.In the development of home smart cloud audio system,it can be divided into embedded platform and cloud platform according to different platforms.First,the embedded platform includes the development of embedded underlying functions and the design and implementation of Android client.The underlying functions of embedded system mainly include infrared remote control,U Disk,HDMI display,Bluetooth function according to peripheral Bluetooth module,and Linux driver development according to peripheral FM radio module.Android client is based on Android system to develop user visual interface and complete a series of function development,including visual operation of embedded underlying functions,broadcast control of media resources such as network radio?video and clips,user rating upload and so on.Secondly,the development of cloud platform includes the design and implementation of information management subsystem and recommendation service.The information management subsystem mainly includes Android resource access,system user login and logout management,user management,network resource management,personalized recommendation management and user rating management.The recommendation service is dynamically deployed in the server in file plug-in mode.By reading the user rating records in the database and using the proposed improved user similarity algorithm as the core step,the personalized recommendation of users is realized,and the hot list recommendation is used as the user cold start solution.This thesis will design the above content in detail through two chapters and show the implementation effect.