Intelligent customer service question answering method and device based on text similarity and storage medium

Abstract

The invention relates to the technical field of artificial intelligence, and discloses an intelligent customer service question answering method based on text similarity, which comprises the followingsteps of collecting all questions and answers corresponding to the questions in a current scene, carrying out text word segmentation and keyword extraction on the questions, and constructing a text corpus of the current question answering; constructing a convolutional neural network model structure based on the similarity between the text keywords; using the text corpus and the convolutional neural network model to train the convolutional neural network model; and receiving a question inputted by a user, inputting the question into the trained convolutional neural network model, predicting the question in the corpus with the highest text similarity with the inputted question, and sending an answer corresponding to the question to the user. The invention further provides an intelligent customer service question answering device and a computer readable storage medium. According to the invention, the accuracy and the calculation speed of intelligent customer service question answering can be ensured.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, in particular to an intelligent customer service question answering method, device and computer-readable storage medium based on text similarity. Background technique [0002] The intelligent customer service system is deeply loved by customers because of its flexible user interaction characteristics that can meet the needs of people to obtain answers. In the intelligent customer service question answering system, the similarity calculation of questions is of great significance. Inquiry, by judging that the questions are similar, retrieving similar questions in the corpus, and recommending the answers to customers, so that customers can get answers to questions more quickly. [0003] Since in most scenarios, the customer's questions are usually composed of topics and descriptions, the length of the customer's questions varies, and the questions will contain a lot of backgr...

AI Technical Summary

Problems solved by technology

[0003] Since in most scenarios, the customer's questions are usually composed of topics and descriptions, the length of the customer's questions varies, and the questions will contain a lot of background information that interferes with the model's judgment of whether the questions are similar. Therefore, simply using cosine similarity, Jaccard Using distance, VSM and other methods to establish an intelligent customer service system, the accuracy rate obtained will be very low, resulting in the answers obtained may not necessarily be the answers to the questions customers want

Images