Title :
Oil Refining Enterprise Performance Evaluation Based on DEA and SVM
Author :
Song, Jiekun ; Zhang, Zaixu
Author_Institution :
Colledge of Econ. & Manage., China Univ. of Pet., Dongying
Abstract :
Enterprise performance evaluation is an important means of enterprise management, which can diagnose the whole development status of enterprise. Data envelopment analysis (DEA) is one of the most frequently used evaluation methods and support vector machine (SVM) is a novel method of data mining, which can be used for prediction and regression. Based on DEA and SVM, the paper proposes a method for evaluating and predicting enterprise performance. First, DEA method is used to evaluate DEA efficiency of all the oil refining enterprises performance. Then the input/output data and results of some decision making units (DMUs) are selected as the learning examples to train the SVM network and the others are used as the test examples to test the network. If the SVM network is testified well, a synthetic evaluation formula can be given to predict the DEA efficiency of a new DMU. A real example testifies the efficiency, practicability and intellectual ability of this method.
Keywords :
data envelopment analysis; data mining; decision making; learning (artificial intelligence); oil refining; regression analysis; support vector machines; DEA; SVM; data envelopment analysis; data mining; decision making unit; enterprise management; machine learning; oil refining enterprise performance evaluation; regression analysis; support vector machine; Data envelopment analysis; Data mining; Hydrogen; Knowledge management; Lagrangian functions; Oil refineries; Quadratic programming; Risk management; Support vector machines; Testing; data envelopment analysis; oil refining enterprise; performance evaluation; prediction; support vector machine;
Conference_Titel :
Knowledge Discovery and Data Mining, 2009. WKDD 2009. Second International Workshop on
Conference_Location :
Moscow
Print_ISBN :
978-0-7695-3543-2
DOI :
10.1109/WKDD.2009.43
