Unbalanced concept drift data flow classification method based on G-mean dynamic weighting

Abstract

The invention discloses an unbalanced concept drift data stream classification method based on G-mean dynamic weighting, and belongs to the field of data stream classification. According to the method, through an online weighting mechanism, the weights of all base classifiers are updated once and are not influenced by class distribution when a new instance rather than a complete data block arrives, the creation time of the classifiers and the Gmean performance of the classifiers on the latest p data are considered when the base classifiers are updated, and therefore the classification accuracy of the base classifiers is improved. The G-mean performance is insensitive to the class distribution of the data, the importance of multiple classes and few classes can be balanced, the classification accuracy of the minority classes is improved, meanwhile, the TPR and the TNP are incrementally calculated through the time attenuation factor, and any historical data does not need to be stored; meanwhile, two elimination mechanisms are used for controlling the scale of the integrated model, and the high efficiency and accuracy of decision making are kept.

Description

technical field [0001] The invention relates to a method for classifying unbalanced concept drift data streams based on G-mean dynamic weighting, and belongs to the field of data stream classification. Background technique [0002] In the era of big data, the explosive growth of information has led to the widespread appearance of data streams in various fields, such as wireless sensor data streams, bank transaction data streams, etc. How to realize the accurate classification of data flow is an inevitable problem to be solved; compared with the classification of traditional static data, there are the following problems in the classification of data flow: [0003] On the one hand, the underlying distribution or target concept of data in a data stream changes over time, a phenomenon commonly referred to as concept drift. Concept drift will lead to a significant drop in the performance of classification models trained on past data, and the classification accuracy rate will be ...

AI Technical Summary

Problems solved by technology

[0006] In order to solve the problem of concept drift existing in the current classification of data streams and the problem of low classification accuracy for minority classes, for binary classification data streams, the present invention introduces an online update mechanism based on the data block integration method, and proposes a OGUEIL, an online unbalanced data stream classification method based on G-mean weighting, is based on the integrated framework. Every time a new instance arrives, each base classifier and its weight are incrementally updated, and the few-class instances are randomly over-sampled. Save historical data and periodically add new candidate classifiers

Images