Data mining method and device and computer readable storage medium

Abstract

The embodiment of the invention discloses a data mining method and device and a computer readable storage medium. The embodiment of the invention comprises the following steps of extracting features on a to-be-processed data set; constructing a feature space, extracting node features from the feature space; generating graph data of the to-be-processed data set, wherein the graph data at least comprises one node, screening out a data cluster corresponding to the node from the graph data, calculating data purity of the data cluster to obtain intra-cluster purity of the data cluster, and when theintra-cluster purity is lower than a preset purity threshold, obtaining corresponding data of the node in a to-be-processed data set to obtain mined data. According to the scheme, all feature information in the data cluster is investigated. A bad file is evaluated through the purity in the data cluster, then bad file mining is carried out, the transition dependence on feature representation is reduced, the bad file in the data can be mined more quickly, efficiently and accurately, and therefore the hit rate of the bad file in the data is increased.

Description

technical field [0001] The present invention relates to the field of communication technology, in particular to a data mining method, device and computer-readable storage medium. Background technique [0002] In data mining scenarios, both image data and text data require pure data, but limited by the representation ability of the model, the data generated by classification, clustering, etc. often cannot guarantee the in-cluster data due to its bad case. Purity, in the prior art, the purity of the data cluster is judged by manual search and simple calculation of the distance between two features in the data cluster. [0003] In the process of research and practice of the prior art, the inventors of the present invention found that a large amount of labor cost is consumed by manual search, and the simple calculation of the distance between two features in the data cluster is due to the diversity and difference of the data clusters. Large, making it difficult to distinguish b...

AI Technical Summary

Problems solved by technology

[0003] In the process of research and practice of the prior art, the inventors of the present invention found that a large amount of labor cost is consumed by manual search, and the simple calculation of the distance between two features in the data cluster is due to the diversity and difference of the data clusters. Large, making it difficult to distinguish between bad files and normal data, thus resulting in a low hit rate for Bad cases

Images