一种改进的决策树遗传算法融合模型  

An Improved Model Based on Decision Tree and Genetic Algorithm

作　　者：柳炳祥[1] 曹坤[1] 洪晶[1] 

机构地区：[1]景德镇陶瓷学院信息工程学院,景德镇333403

出　　处：《微计算机信息》2007年第04X期258-259,160,共3页Control & Automation

基　　金：江西省教育厅科研项目资助(04JJ23)

摘　　要：针对目前遗传算法存在随机性大、收敛速度慢等缺点,文章提出了一种改进的决策树遗传算法融合模型。该分类模型充分发挥了遗传算法的优点,有效提高分类的准确度,将模型应用到电信行业客户流失分析中,验证了该算法的可行性和有效性。There are some shortcomings in genetic algorithm, such as big randomness, slow velocity of convergence. Then a classification algorithm, which is combined by decision tree and improved genetic algorithm, is put forward. The algorithm fully displays the excellence of genetic algorithm and efficiently improves the accuracy of classification. At last this algorithm is applied to the telecommunications industry customer loss analysis to confirm the efficient of algorithm.

关 键 词：决策树 遗传算法 局部补差算子 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]