Abstract: The complex structure in enterprises has a pivotal effect on enterprise overall performance. In this paper, performance evaluation for enterprises with parallel structure is studied. An integrated model with support vector machine (SVM) classifier and data envelopment analysis (DEA) is proposed. First, the DEA efficiency for each internal department in enterprises is analyzed and used as classification criterion and training data. Then, SVM classifier is applied to classify all departments based on class criterion. Finally, the classification information is integrated into the parallel DEA models so that the overall performance efficiency for enterprises with parallel structure is figured out. With an empirical analysis, the practicality, effectiveness, and operability are testified by using the integrated model. This model provides a referential analytic tool for measuring the enterprises with parallel structure.

Key words: SVM classifier, parallel chain structure, data envelopment analysis(DH