Electric charge notification and customer appeal collection classification method and system based on NLP technology, and storage medium

Abstract

The invention belongs to an electric power data processing technology, and relates to an electric charge notification and customer appeal collection classification method and an electric charge notification and customer appeal collection classification system based on an NLP technology, and a storage medium. The electric charge notification and customer appeal collection classification method comprises the steps of: sorting a classification knowledge graph, establishing a classification framework meeting actual demands, carrying out classification and marking of electric charge customer service work orders manually, and extracting a classification rule through iterative operation; carrying out text cleaning on the electric charge customer service work orders to be classified; constructinga professional lexicon to perform text word segmentation on the cleaned electric charge customer service work orders; using a TF-IDF algorithm to represent text vectors of the electric charge customerservice work orders; screening effective features of the text vectors by adopting an information gain method; and adopting an SVM support vector machine algorithm to classify the work orders. According to the electric charge notification and customer appeal collection classification method and the electric charge notification and customer appeal collection classification system, the complaint content can be subjected to precise word segmentation, the optimal classification of the text is calculated, the automatic classification of the customer complaint content in customer service work ordersof customer service electric charge types is realized, and the problems that the granularity of manual classification is not fine enough, the resource consumption is excessive, the classification speed is lagged and the like are solved.

Description

technical field

[0001] The invention belongs to the technical field of electric power data processing, relates to machine learning, NLP and customer work order classification, and specifically relates to a classification method, system and storage medium for electricity bill notification and collection of customer appeals based on NLP technology. Background technique

[0002] In order to conscientiously implement the deployment of marketing work and further promote the application of marketing big data, customer-centric, based on the basic data of 95598 electricity bill work orders, use natural language processing, machine learning and other big data processing technologies to build customers' notification and reminder of electricity bills The classification model of appeals focuses on customers’ appeals for electricity bill notifications and feedback on electricity bill payment in real time, optimizes the content of electricity bill notifications and achieves group different...

Images