Hierarchical clustering method and device

Abstract

The disclosure relates to a method and device for hierarchical clustering, belonging to the field of data mining. The method includes: obtaining a set of data objects to be clustered, the set of data objects includes multiple classes, each class corresponds to at least one data object; clustering the data objects corresponding to the first class to obtain a clustering result, the The number of data objects corresponding to the first category exceeds the first preset threshold, the clustering result includes a plurality of clusters, and each cluster includes at least one data object; according to the clustering result, the data objects corresponding to the first category are processed Filter to obtain the representative data objects of the first category; calculate the inter-class distance based on the representative data objects of the first category and the corresponding data objects of the second category; perform hierarchical clustering on the data object set based on the inter-category distance. In the present disclosure, by clustering the data objects corresponding to the first category, the calculation amount is reduced, the calculation time and resources are saved, and the clustering result is more reliable, which is beneficial to subsequent data analysis.

Description

technical field [0001] The present disclosure relates to the field of data mining, in particular to a method and device for hierarchical clustering. Background technique [0002] In the field of data mining, it is usually necessary to analyze a large amount of data in order to obtain valuable analysis results. Clustering algorithm is an important algorithm for analyzing data in the field of data mining. This algorithm is used to classify a set of multiple data according to different categories of data. aggregated into one category to facilitate subsequent data analysis. Among them, hierarchical clustering is a commonly used clustering algorithm. [0003] When the related technology implements the method of hierarchical clustering, it calculates the distance between two classes, that is, the inter-class distance, so as to merge two classes whose inter-class distance is less than a certain value into a new class. Since each class may contain more than one data object, when ...

AI Technical Summary

Problems solved by technology

[0005] In related technologies, when implementing hierarchical clustering by calculating the distance between classes, the amount of calculation is too large. When there are many data objects contained in a class, it will consume too much time and resources. Moreover, because each class may contain Data objects that do not belong to this class, that is, noise. After using this data object to calculate the inter-class distance and form a new class, more noise may be introduced, resulting in poor clustering results, which is not conducive to subsequent data analysis

Images