Machine learning-based customer value model optimization method

• Main Authors: TANG ZIYU, LI WEI, LI XINGLONG
• Format: Patent
• Language: Chinese; English
• Published: 19.01.2018
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; PHYSICS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

The invention relates to a machine learning-based customer value model optimization method. The method comprises the following steps of 1: extracting customer value model data of N customer subjects in different periods through a random sampling method to obtain initial model data samples Si (i=1,2,3...N); 2: correspondingly training the initial model data samples Si (i=1,2,3...n) by using a bagging machine learning method to obtain N independent individual weak learners Hi (i=1,2,3...N); 3: combining the individual weak learners Hi (i=1,2,3...N) into a strong learner H through a stacking combination policy; and 4: taking the strong learner H as an optimal model rule, and inputting current customer value model data samples to the strong learner H, wherein a result obtained by the strong learner H is an optimal result model. 本发明涉及种基于机器学习的客户价值模型优化方法，包括如下的步骤：步骤1：通过随机采样法提取N个客户主体不同时期的客户价值模型数据，得到初始模型数据样本Si(i＝1、2、3...N)；步骤2：对给个初始模型数据样本Si(i＝1、2、3...n)分别使用bagging机器学习方法，相对应地训练出N个独立的个体弱学习器Hi(i＝1、2、3...N)；步骤3：通过stacking结