Research And Application Of Data Privacy And Service Recommendation Algorithm In Smart City

With the continuous increase of urban population and the continuous expansion of urban scale,cities have played a more and more important role in the development of human society.Therefore,smart city was born.Its emergence helps to strengthen urban management and improve people’s quality of life.With the development of smart city,there are many problems in the city,such as urban traffic jam.The occurrence of traffic jam will bring huge losses to the city every year and affect the efficiency of urban operation.a large number of services are used all the time in the city.Due to the gradual emergence of the same or similar services,how to recommend high-quality services to users has become a challenging task.In addition,all current service recommendation methods need to predict Qo S based on users’ historical data.However,users’ historical information contains a lot of users’ privacy information.In addition,users’ calls to services are often affected by factors such as time and space.How to integrate these factors affecting the accuracy of recommendation into the process of service recommendation is also a big problem.To solve these problems,this paper proposes a time aware service recommendation framework that can protect privacy.The specific contents are as follows:(1)In order to solve the problem of privacy information disclosure contained in user history data,this paper proposes a service recommendation framework with privacy protection function.Using the method of differential privacy to add noise when collecting user data,and then use the service quality data after adding noise for prediction,which can effectively protect the user’s privacy information.(2)Considering that the traditional service recommendation methods based on Qo S prediction process data from a single dimension,without considering the diversity of data.This paper uses the method of multi-dimensional segmentation to process the data,forecasts from the three dimensions of user,service and time,and then uses the weighted method to make the final prediction.The transformation from single dimension to multi-dimensional prediction improves the application scope of service recommendation algorithm,so that it can better serve users.(3)Considering that the problems of sparse data and cold start in service recommendation may lead to the decline of the accuracy of service recommendation,the L1 normal form low rank matrix decomposition can be predicted even when the amount of initial data is very small.Before Qo S prediction,K-means clustering algorithm is used to divide the data into multiple clusters.L1 normal form low rank matrix decomposition for each cluster can speed up the execution speed of the algorithm and improve the accuracy of Qo S prediction.(4)According to the time aware service recommendation framework with privacy protection proposed above,this paper realizes the road recommendation system based on Qo S prediction in smart city.The system is mainly used to solve the problem of road congestion in smart cities.It can intelligently recommend roads for users and prevent road congestion.