Abstract: When decision tree of futures forecasting is constructed by the method of current C4.5 algorithm, there exist drawbacks about low forecast accuracy, and prediction model is utilized difficultly. Therefore, it creates an C4.5-[K] algorithm oriented to futures data, in which a new parameter [K] is introduced, property metric information gain ratio is adjusted, and prediction model of decision tree is built for prediction in C4.5 algorithm. Experimental results demonstrate that the improved algorithm can enhance the capacity for futures forecasting effectively.

Key words: C4. 5 algorithm, decision tree, futures prediction, data mining

摘要： 在利用现有C4.5算法构建期货预测决策树时，往往出现预测准确率低的弊端，导致预测模型很难使用，为此提出了一种面向期货数据的C4.5-[K]算法。该算法的主要思想是通过在C4.5算法中引进新的参数[K]，调整属性度量标准信息增益率的取值范围，进而构建决策树预测模型进行预测。通过实验表明，该改进算法能有效提高期货预测能力。

关键词: C4.5算法, 决策树, 期货预测, 数据挖掘