Data clustering method and apparatus, computer readable medium and electronic device

Abstract

The invention provides a data clustering method and apparatus, a computer readable medium and an electronic device. The data clustering method comprises the steps of obtaining a to-be-clustered data set; calculating a distance between each piece of data in the data set and a clustering center of each existing type; if the distance between any data in the data set and the clustering center of any existing type is smaller than or equal to a distance threshold, including any data in any type; if the distances between any data in the data set and the clustering centers of all existing types are all greater than the distance threshold, creating a new type, and including any data in the new type. According to the technical scheme, during data clustering, a clustering number and the clustering centers do not need to be specified in advance, so that the bad influence on a final clustering result due to wrong selection of an initial clustering center is avoided and the time for the data clustering process can be shortened.

Description

technical field [0001] The present invention relates to the technical field of data processing, in particular to a data clustering method, device, computer readable medium and electronic equipment. Background technique [0002] In the process of building the user portrait label model, after extracting user features and standardizing the feature data, there are many scenarios for building labels based on clustering, such as promotion sensitivity clustering, comment sensitivity clustering, user loyalty clustering, etc. . Clustering is to divide the user set into different classes or clusters according to a certain standard under the corresponding user characteristics, so that the similarity or distance of the user characteristics in the same class or cluster is as large as possible or the distance is as small as possible, and at the same time they are not in the same class. Or the difference of user characteristics in the cluster is also as large as possible. In short, after...

AI Technical Summary

Problems solved by technology

[0006] (2) In the k-means algorithm, the initial clustering center needs to be determined artificially. Different initial clustering centers may lead to completely different clustering results. Once the initial value is not selected properly, effective clustering results may not be obtained. ;

[0007] (3) The k-means algorithm is sensitive to outliers and cannot detect outliers, and outliers sometimes have a great impact on the accuracy of the cluster center;

[0008] (4) The k-means algorithm needs to continuously adjust the sample classification and continuously calculate the adjusted new clustering center, the convergence is slow and the clustering time complexity is high. When the amount of data is very large, the time overhead of the algorithm Very big;

[0009] (5) The k-means algorithm needs to scan the full amount of data multiple times, and cannot cluster real-time data

Images