A Novel Approach Based On SVM To Forecasting The Thermal Power Plant Construction Cost

United the thorey of Statistics and Machine learning and Optimized theory, SVM establishes the automatic study model according to the structure risk minimum principle the structure.It changes the model structure by regulating control parameter.It is better than the neural network and fuzzy identification in non-linearity, high dimensions recognition and partial minimum point and so on.Recently,it has been used in many aspects,like the pattern recognition, the regression analysis and the characteristic extraction and so on. This paper try to use SVM to forecast the Thermal Power Plant Construction cost. It will utilize the svm system, combine the thermal power engineering projects tendency and the characteristic, establish the forecast model. Last, to proove SVM forecasting model is usefull in thermoelectric power station project construction cost and is mean for the practical.