Research On The Enterprise's Core Competitiveness Evaluation Of High-tech Companies Based On Improved BP Neural Network

For the high-tech enterprises, the strong core competitiveness to gain competitive advantage is its fundamental element. And the research about core competence theory of enterprises is also a hot issue in the theory community and the corporate governance practices now. But it has been the lack of systematic quantitative and effective theoretical identification and evaluation method in the high-tech enterprise's core competitiveness.Based on the summary of the core competitiveness theory and the high-tech enterprises theory, first, this article analyzed the characteristics and its core competitiveness of the unique nature and its composition of the high-tech enterprises as a special case. Second, analyzed the existing evaluation methods of the core competitiveness,then get the use of BP neural network model to evaluate the core competitiveness of enterprises. On the view of many scholars at home and abroad about high-tech features of core competence, in the consider of the availability, quantitatively and the representative of indicators, building a index system which contain 4 evaluation factors,20 evaluation indexes of high-tech enterprise core competitiveness evaluation. Then, to the BP neural network convergence speed is slow, less easy to fall into local minimum, Levenberg-Marquardt algorithm is used to solve these problems; designed a 20-40-1 network model, and used the MATLAB software and its Neural Network Toolbox to achieve the evaluation process. Finally 10 high-tech enterprises in different areas of Henan Province were used as the neural network's learning and training indicator data, using 2 enterprises'assessment data to simulate and test, the result shows that the evaluation index system and improved BP neural network model application has good practicability. In addition, we discussed the evaluation method of horizontal and vertical extension of applications, and also the core competitiveness identification of high-tech enterprises.In this study, by using BP neural network trained and tested, proved that the improved BP neural network quick learning, and higher precision, can be a good network model in solving complex system and nonlinear nature. Avoid excessive disturbance of subjective factors in the evaluation process, and has a good performance in evaluating the core competitiveness of high-tech enterprises. This study provides may has some theoretical and practical value for the enterprises to evaluate their own core competence accurately, validly and quantitatively, then understanding their market position objectively, so that appropriate management strategies for competitive advantage properly, and finally get best economic benefits.