A Study On Performance Prediction On Integrated Support Vector Machines Of Hi-Tech Enterprise

Independent innovation capability is the core competitiveness of High-tech enterprises. Now that competition among enterprises is more and more intense, technology has developed more and more rapidly and under all those situation, development of High-tech enterprises are facing greater challenges and future development is uncertain, especially those high-tech enterprises listed to GEM. Since listing on the GEM board is easier compared to listing on the motherboard, enterprises development listing on the GEM board is uncertain. So we need an effective method to help shareholders, creditors and some other shareholders identify and predict enterprises which are of good performance and help those shareholders to make decision in order to get the most earnings under risk. SVM(Support Vector Machines) is a new way of machine learning and it is applied to small sample, nonlinear and high dimension classification problems. In the paper, firstly we review the research of other researchers about business performance predicting, then we use high-tech enterprises listing on the GEM board as a sample, building performance prediction model using Support Vector Machines. Besides, we also use Support Vector Machine ensemble forecasting model. Then we use the sample to test the model and get the accuracy of prediction.The results of the research shows that prediction accuracy of Support Vector Machines forecasting model is lower than Support Vector Machine ensemble forecasting model. Besides, the closer to forecast year, the accuracy of prediction is higher. The model we build in the paper can help shareholders make investment decision.