Design And Implementation Of A Logistics Distribution Service Recommendation System Based On Spark

With the rapid development of e-commerce,the market's demand for logistics and distribution services has increased rapidly,and the logistics and distribution industry has grown unprecedentedly.At present,there are many distribution services provided by logistics providers.While these services meet the basic delivery requirements of customers,the quality of service is not the same.The customer's demand for service costs,timeliness,security and other factors tend to be individual due to individual differences.As a result,the number of distribution services that each customer is actually interested in is limited,and it takes a lot of time to find a logistics distribution service that suits them.So howAllowing customers to easily and accurately obtain logistics distribution services that meet individual needs has become an important research issue in modern logistics and distribution industry.The above problems can be solved by constructing a recommendation system.In order to make the recommendation system have good big data processing capability,this paper combines the recommendation system with the Spark big data calculation framework and applies it to the field of logistics and distribution to help customers choose the logistics distribution service that meets individual needs.This completes the following tasks as a starting point:(1)In order to obtain the customer's personalized needs and interests for logistics and distribution services,establish a customer preference model,describe the customer's needs and interests from multiple dimensions,and use the space vector representation to analyze the collected static preference information and dynamic preference information.And matrix representation to achieve customer preference modeling.(2)In order to obtain the characteristics of logistics and distribution services,a service evaluation model was established.By collecting the user's score information on multiple attributes of service quality,a comprehensive score was obtained for each attribute of the service.The weights corresponding to each attribute of the service are obtained,and the score value of each attribute is weighted and summed to obtain the comprehensive score of the service.On this basis,the distribution service is vectorized to realize service feature modeling.(3)Combine the customer's preference model with the service feature model to study the traditional recommendation algorithm.According to different applicable situations,propose a service recommendation algorithm based on multi-attribute collaborative filtering,content-based,and demographic-based services.Analyze and verify the effectiveness of each algorithm in different situations.(4)Based on the research of logistics distribution service recommendation algorithm and demand analysis,design a logistics distribution service recommendation system from the aspects of architecture,function,and database,etc.,and determine the use of Spark computing framework to develop recommended modules to complete related platforms.The environment is set up to implement the logistics distribution service recommendation system on the Spark platform.