Unsatisfaction reason tracing method based on improved clustering algorithm

Abstract

The invention provides an unsatisfaction reason tracing method based on an improved clustering algorithm in the technical field of computers, and the method comprises the steps: S10, obtaining a large amount of unsatisfaction data, and extracting the feature data of each piece of unsatisfaction data, so as to construct a first feature data set; s20, removing abnormal data in the first feature data set to obtain a second feature data set; s30, selecting an initial clustering center feature based on the second feature data set; step S40, creating a clustering model based on a kmeans clustering algorithm and the initial clustering center features, and clustering the second feature data set by using the clustering model to obtain a plurality of clusters; step S50, screening out a strong discrimination feature from each cluster; step S60, clustering the to-be-traced data to a nearest cluster by using the clustering model; and step S70, tracing the dissatisfaction reason from the to-be-traced data based on the strong discrimination feature and the cluster. The method has the advantages that the dissatisfaction reason tracing efficiency and accuracy are greatly improved.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a method for tracing the source of dissatisfaction based on an improved clustering algorithm. Background technique [0002] At present, the competition among various communication operators is very fierce. It is already difficult to gain a competitive advantage only by relying on unilateral reforms of networks, channels, and prices. This makes the competition gradually change from price competition to service competition. Understanding and meeting customer needs, Improving customer satisfaction is the key for communication companies to gain competitiveness. [0003] In order to enhance their competitiveness, traditionally, communication companies conduct satisfaction surveys on a regular basis to understand the current customer service evaluations of communication companies through sampling surveys, and then conduct traceability analysis on all aspects of dissatisfied users to ...

AI Technical Summary

Problems solved by technology

[0004] Because the traditional method requires manual intervention, the reasons for user dissatisfaction are divided into large categories based on manual experience, and then according to the actual behavior of users, the manual classification of dissatisfied users requires a lot of labor costs. In addition, the artificially divided dissatisfaction categories have incomplete coverage, which affects the accuracy of the source of dissatisfaction.

Images