Clustering evaluation improvement method and device, computer equipment and storage medium

Abstract

The invention discloses a clustering evaluation improvement method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining a clustering sample of an advertisement business user, and constructing a clustering sample cluster according to the clustering sample; calculating distance values between the clustering samples and other clustering samples in the clustering sample cluster respectively, and forming distance vectors; normalizing the distance vector to obtain a weight vector and a weighted average value; obtaining a weighted average value of distances between the clustering samples and other clustering samples in the clustering sample cluster as a cohesion degree in the cluster; taking the weighted average value between the clustering samples and all clustering samples in the clustering sample cluster with the minimum distance as the separation degree of the clustering sample cluster with the minimum distance; and taking the difference between the separation degree and the cohesion degree as a molecule, taking the maximum value of the cohesion degree and the separation degree as a denominator, and taking a score value as an evaluation coefficient. The classification of advertisement service users is more reasonable, the user identification is improved, and the average weight effect of arithmetic mean values is improved.

Description

technical field

[0001] The invention relates to the technical field of advertisement data processing, in particular to an improved method, device, computer equipment and storage medium for cluster evaluation. Background technique

[0002] In the current technology, in the advertising business, it is necessary to classify user groups to facilitate accurate user marketing and operations. The accuracy of classification determines the quality of the effect. The k-means algorithm is commonly used for clustering. However, when k-means clustering is currently performed on advertising business users, the cluster evaluation adopts the silhouette coefficient, which uses the average distance between the advertising business user sample x and other points in the cluster as the cohesion in the cluster Degree a, the average value between the sample x and all points in the nearest cluster is regarded as the degree of separation b from the nearest cluster, and then the difference between th...

Images