摘要
为了有效提高电力企业客户满意度及主动服务意识,结合电力客服工单文本特征,构建了电力客服工单情感分析模型。先通过TF-IDF思想进行工单关键词提取,采用word2vec训练得出每个词语的词向量,通过计算余弦相似度将高相似领域词汇扩充到情感词典,再进行工单倾向性分析及文本分类。通过实验分析验证该方法的有效性,实验结果表明,相较于原始情感词典,进行词典扩充及工单情感倾向性分析方法更具优势,准确率更高,可为电力企业客户关系管理提供一定的参考。
In order to improve the customer satisfaction and active service consciousness of the electric power enterprises effectively,the textual characteristic of the power customer service order is combined to construct the sentiment analysis model of the power customer service order. The keywords of the service order are extracted according to TF-IDF thought. The word2 vec training is used to get the word vector of each word. The cosine similarity is calculated to expand the high similarity field vocabulary to the sentiment dictionary. The service order sentiment analysis and text classification are performed. The validity of the method is verified with experimental analysis. The results show that,in comparison with the original sentiment dictionary,the method of dictionary expansion and service order sentiment tendency analysis is superior,has higher accuracy,and can provide a certain reference significance for the customer relation management of power enterprise.
出处
《现代电子技术》
北大核心
2017年第11期163-166,171,共5页
Modern Electronics Technique
关键词
情感分析
情感倾向性
词典扩充
电力客服工单
主动服务
sentiment analysis
sentiment tendency
dictionary expansion
power customer service order
active service