Automatic label labeling method and system, equipment and storage medium

Abstract

The invention discloses an automatic label labeling method and system, and the method comprises the following steps: a word grouping step: building a plurality of independent word banks based on business corpora, and carrying out the word grouping of the words of each word bank; a label and label rule definition step: on the basis of each word bank, after label groups are defined according to business elements, selecting or customizing words from the word banks as labels, and defining label rules on the basis of a label automatic labeling matching mode; and a label labeling step: based on a label rule, matching the words in the to-be-labeled text with the words in the word groups, labeling the words in the to-be-labeled text meeting a matching condition with labels, and storing an automatic label labeling result into a label result table to complete automatic labeling of the labels. The invention provides a set of automatic label labeling method, which is used for automatically marking through an algorithm model according to a word bank, a label and a rule system corresponding to industry maintenance, and providing evaluation step-by-step training to improve the labeling accuracy.

Description

technical field [0001] The present application relates to the field of data analysis, in particular to an automatic labeling method, system, computer equipment and computer-readable storage medium. Background technique [0002] Currently, in offline service scenarios, service quality is difficult to monitor and measure due to the uncertainty of the scenarios. We can concretize the dialogue in the service process into text, and the service quality into corresponding labels. Through automatic and accurate marking of texts, we can effectively extract the labeling rules of customers in different scenarios and different speech themes, and provide sales services The analysis model provides data support and provides digital and visual basis for customer supervision. [0003] Manually evaluate the text labels from the table offline, and calculate the evaluation results, label coverage and accuracy by itself. [0004] At present, there are the following bottlenecks in related techn...

AI Technical Summary

Problems solved by technology

[0005] (1) The labor cost is high, and the division of labor and collaboration cannot be carried out effectively

[0006] (2) The accuracy of the evaluation cannot be guaranteed, and the diversity of tags is difficult to accurately compare right and wrong manually

[0007] (3) The calculation method of coverage and accuracy is difficult to calculate manually

[0008] (4) The "cold start" stage that cannot solve the lack of awareness of new business scenarios

Images