The Research And Application Of Equipment Management System Based On Decision Tree

The traditional equipment management system is primarily designed for business, including maintenance, query and statistics analysis functions of equipment file, operation, repairing and maintenance data, but lack of decision support function. The system has accumulated a large amount of operating and maintenance data, and the manager's decision-making mainly depends on the result of statistics and analysis, costing a lot of time. With the development of Information technology, Data Mining has become a powerful tool for decision support. As an important technology in the field of data mining, Decision Tree is rarely applied in the operation management of equipment in oilfield. So putting the Decision Tree into equipment management system, obtaining the information based on rules to guide decision-making and forecasting, has an important theoretical research and application value.Firstly, the article describes the current research situation of the Decision Tree and Equipment Management Systems home and abroad. Combining with the actual situation of equipment management in oilfield, the article describes an equipment management system design based on Decision Tree, in which it proposes a model for equipment operational state diagnosis. The decision tree algorithm is used in the model, which analyzes the key indicators such as good rate, using rate, falult rate and so on, derives the rule information based on the knowledge. Also the model is a general model for the similar machinery equipment in oilfield and can achieve rapid diagnosis for equipments.Secondly, by comparing the classification algorithm of decision tree, ID3 is selected as main algorithm for self-learning module. For the performance of data processing and property values bias problem, the article proposes an improved algorithm, which introduces Taylor Formula and Attribute Similarity algorithm, aiming at simplifying the information entropy operation and overcoming the problem of attribute selection dependence. And the improved algorithm has better versatility.Lastly, the article implements the previous and the improved algorithm by JAVA language. On the basis of the samples of oilfield equipment data, the verification experiment is also carried out. The result shows that the improved algorithm avoids the shortcomings of the previous algorithm and improves the generating performance and classification accuracy of decision tree.