基于机器学习模型的客服短文本分类技术研究

Research on short text classification technology of customer service based on machine learning model

客服的本质工作是对用户登记的问题进行分类,并根据分类结果将其转发对应的部门进行处理,分类通常依赖人工分析,处理效率较低。为了提高客服人员的工作效率,对基于机器学习模型帮助客服自动分类用户反映问题的方法进行研究,准确进行文本分类和识别。研究选取某公司登记的用户问题数据,分别采用哈希向量化(HashVectorizer)和词频-逆文档频率(TF-IDF)等技术构建文本向量,对比分析多种机器分类模型,选取最优模型,且取得较好分类效果。

The essence of customer service is to classify the user registration problems and forward them to the corresponding departments for processing according to the classification results.Classification usually relies on manual analysis,and the process-ing efficiency is low.In order to improve the work efficiency of customer service personnel,a method based on machine learning model to help customer service automatically classify users’questions,and accurately classify and recognize text.Research and se-lect the user problem data registered by a company,constructs the text vector using HashVectorizer and TF-IDF technologies re-spectively,compares and analyzes various machine classification models,selects the best model,and achieves good classification results.