关键词: 支持向量机, 分类, 回归, 产量预测

Abstract: For non-linear problem, the forecasting technique of pre-classification and later regression was proposed, based on the classification approach of Support Vector Machine (SVM). According to the actual requirements and professional knowledge, the sample cluster was classified first to decide the types of the test samples. Next the values of the test samples were forecast with the regression algorithm. Compared with other forecasting techniques and their forecasting results, this algorithm outperforms others in grain output prediction. The findings of the experiment show that the fitted value obtained from the forecasting technique of pre-classification and later regression is much more accurate than that from regression.

Key words: classification, regression, production forecast