Forecasting the Trend of Construction Cost Indices for Taiwan with Employing Support Vector Machine

• Main Authors: Yu-Ying Chang, 張玉瑛
• Other Authors: Yi-Shuo Huang
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/09462437913330103590

碩士 === 朝陽科技大學 === 營建工程系碩士班 === 95 === Construction Cost Indices (CCI), one of the economic indicators for Taiwan''s construction industry, is a subsequent numerical statistics. If the trend of the CCI can be forecasted, the risk of price fluctuations in the construction industry can be reduced positively. In this study, support vector machine (SVM) is employed to predict the trend of CCI. From the results, the developed SVM model demonstrates that the method does correctly predict the tendency of CCI in a specified interval with the limited information. Those variables affecting the economic environment are also included in the developed SVM model such that the prediction retains the validity when the economic environment changes. In the thesis, we found out that with the limited samples, the model made with SVM can correctly fit the actual index trend. The results shown in the thesis demonstrate the applicability and accuracy of the developed approach used in this study.