一种基于回归树模型的精准营销业务解析方法

A Precision-marketing Business Analysis Method Based on Regression Tree Model

数字营销是各大电信运营商进行客户价值挖掘及存量维系的总趋势,当前业务人员均积极使用各种机器学习模型对目标客户群进行圈选以支撑后续的深度精准营销工作,但是不易理解的“黑盒”化机器学习模型给业务人员的营销策略制定带来了巨大的挑战。文章创新性的提出了一种模型可解释的方法,其通过回归树算法对前向的机器学习挖掘模型进行反向解释,从而透析出有明显特征差异的业务规则链条,助力业务人员实现精准营销和精细化决策。

Digital marketing is the general trend of telecom operators in customer value mining and owncrship maintenance,At present,Business Specialist are actively using various machine learning models to circle the target customer groups to support the subsequcnt in-depth and accu-rate marketing work,but the"black box"machine learning model,which is not casy to under-stand,has brought huge challenges to business personnel's marketing strategy formulation.This paper innovatively proposes a model interpretable method,which reversely interprets the for-ward machine learning mining model through the regression tree algorithm,thus dialyzing the business rule chain with obvious characteristics differences,and helping business personnel to a-chieve precision marketing and refined decision-making.The results show that under the same access channels and marketing techniques,the conversion rate of marketing customers using the models reversc explanation ability is significantly better than that of marketing results without the models reverse explanation ability.