Personalized Recommendation Method And System Implementation Of Metrological Industry Services

With the rapid development of cities in recent years,emerging industries continue to appear,and the application field of metrological and inspection industry has become more and more widely.In the traditional metrological service industry,customers choose their own metrological services through their own subjective intention and the manual recommendation of the service personnel.This method is fast and effective when the customer demand is clear.However,when the customer is not clear about their needs,the traditional manual recommendation method seems to be stretched.According to the characteristics of metrological industry customer groups and business,combined with collaborative filtering algorithm and latent semantic model recommendation algorithm,is proposed and designed for metrology metrological industry customer service personalized recommendation system.The main contents of this paper are as follows:1.Aiming at the problem of sparse data in the metrological industry,a personalized recommendation method based on user clustering and collaborative filtering is proposed.The method incorporates the macro information of the customer's industry and metrological services through the existing customer service preference data.The K-means clustering algorithm is used to cluster the customer data to reduce the searching range of nearest neighbors,so as to shorten the computing time of nearest neighbors.Then,the customer service preference matrix and the customer service category preference matrix are constructed to calculate the nearest neighbor set of the target customers.Finally,the personalized recommendation of each client is implemented by using the user based collaborative filtering algorithm.2.According to the metrological industry more implicit preference data type,proposed the personalized method of latent semantic model based on metrological industry.This method is based on the classical latent semantic model,adding the attribute information of metrological industry customers,to measure the importance of each attribute by logistic regression classification algorithm,according to the attribute information of customers to find similar users.By setting weights,the customer preference value prediction formula is optimized.This method can be recommended by the customer attribute information in customer history data is insufficient,when customer history data more and more,then gradually transition to the use of history data of customers to recommend.By adding customer attribute information,the cold start problem of new customers can be solved,and the scalability of the system can be improved.3.Design and implement a simple personalized recommendation system for Fujian Province metering industry services.According to the needs of customers and administrators,the system needs analysis,and use case diagram to describe the user needs.And the function of the system is designed on the premise of meeting the system requirement.The first part is about the summary of the design,and then on this basis,from two angles of foreground and background to complete the detailed design of the system,a detailed description of the design of the main framework of the system,and gives the corresponding system key module flow chart.Through the PowerDesigner database modeling tool to complete the design of the system database.Finally,the implementation environment of the system and the implementation effect diagram of some main modules are introduced briefly.In this paper,the historical data of customer measurement provided by Fujian Academy of metrology is used as the experimental data set,and the experimental analysis of the proposed method is carried out.The experimental results show that the recommendation method proposed in this paper has high recommendation accuracy in the customer history data set of the measurement industry based on sparse data and implicit feedback data.