An electric power customer service work order sentiment quantitative analysis method based on a similarity word sequence matrix

Abstract

The invention discloses an electric power customer service work order sentiment quantitative analysis method based on a similarity word sequence matrix, and relates to an electric power customer service work order analysis method. A traditional sentiment analysis method cannot effectively discriminate sentiment intensity. The method comprises an emotion word expansion association step based on Word2Vec similarity, a multivariate emotion corpus construction step and a similarity word sequence matrix emotion quantization algorithm step, and is characterized by classifying and combing the work orders; cleaning the data, forming an initialized multivariate emotion word bank based on the Baidu word bank, carrying out work order text word segmentation by adopting a reverse maximum matching algorithm; based on the Word2Vec neural network, constructing the positive words, negative words, negative words, degree auxiliary words and the word vectors of the word sequences fused with customer appeal semantics, generating a learning model fusing appeal emotions through machine learning and training, and expanding a part-of-speech corpus based on a part-of-speech affinity-consensus relation, performing the emotion quantitative calculation by adopting a similarity part-of-speech matrix quantization algorithm, thereby completing the emotion quantitative analysis of a customer service work order, and effectively distinguishing emotion intensity differences.

Description

technical field [0001] The invention relates to a method for analyzing electric customer service work orders, in particular to an emotional quantitative analysis method for electric customer service work orders based on a similarity word order matrix. Background technique [0002] With the development of social economy and the continuous deepening of power system reform, power supply companies can only gain market-oriented competitive advantages by adhering to customer-centricity and improving customer satisfaction. As an important channel window for customer communication and communication, 95598 realizes quantitative emotional analysis of customer appeals through deep mining of customer characteristics and emotional information hidden in customer service work orders, which is conducive to quickly understanding the focus of customer attention, and is conducive to according to customer Identifying potential complaining customers with emotional tendency is conducive to suppor...

AI Technical Summary

Problems solved by technology

[0005] Traditional sentiment analysis methods are based on bag-of-word features and word frequency statistics. At the same time, most studies focus on the classification of sentiment orientation. This method has the following three defects: (1) The context order and semantic understanding of missing words ; (2) Ignoring the semantic difference between words; (3) Focusing on the classification of emotional orientation can not reflect the difference in emotional strength

In short, it cannot effectively identify the intensity of emotion

Images