The Design And Implementation Of Business Tag Management System

With the rapid development of Eleme’s market size and rapid changes of business requirements, the number of restaurants settled in Eleme is increasing and the restaurant tags increased as well, so the restaurant tag specifications also need to meet the changes of market and business. In order to improve the efficiency and accuracy of the restaurant tag classification, and reduce error of manual classification, we developed the restaurant tag management platform based on big data and machine learning techniques.The restaurants settled in Eleme can be classified automatically by the resraurant tag management platform. The platform can adapt to the changes of restaurant menus and update the tag of changing restaurant. All restaurants are classified according to the unified standard of Eleme, this can reflect fairness for all restaurants and avoid restaurants tagging themselves. Tag classification data is the basis of the restaurants, and is the indispensable dimension of search, recommendation and other products. It has important value and significance to the big data statistics and analysis.Eleme Restaurant Tag Management System uses Spring MVC to develop, and the front-end uses jQuery and Bootstrap techniques. It uses MySQL as database and Hive as data warehouse which is used to store offline data to make data analysis and processing easier. As the core of the system, tag smart classification module uses Python to develop, and uses LSI, random forests and other machine learning algorithms for intelligent restaurant classification.At present, the system has been put into a preliminary trial. It will pull data to Hive from online database and compute offline to classify all restaurants intelligently. The reviewed classification data will be pushed to the online database to update. According to statistics, the accuracy rate of classification has reached 90 percent. Later we will continue to optimize the algorithm to get a 95 percent accuracy rate.