Multi-label unbalanced virtual asset data classification method based on integrated learning

Abstract

The invention discloses a multi-label unbalanced virtual asset data classification method based on integrated learning. The method includes the following steps that under a distributed storage framework of virtual assets, firstly, random sampling with replacement is carried out on virtual asset data, then, multi-label data is learned through a feedforward neural network, and the relevance between labels is implied in a trained neural network connection weight; meanwhile, the SMOTE is selected and used for sampling according to the distribution of the labels in sampled data; finally, for improving the generalization performance of a classifier, the integrated learning method is adopted, and the neural network serves a weak classifier in each turn of learning. Compared with the prior art, the classical algorithm Bagging in integrated learning serves as the framework, the feedforward neural network and the SMOTE are integrated in the integrated learning framework according to the characteristics of the unbalanced virtual asset data, and the classification precision can be effectively improved.

Description

technical field [0001] The technology belongs to the field of network and information security, and relates to a multi-label unbalanced virtual asset data classification method based on integrated learning. Background technique [0002] The rapid development of the Internet provides a broad platform for the generation and trading of virtual assets, and promotes the prosperity and development of online transactions. However, both users and providers of virtual asset transactions are faced with the problem of huge and complex virtual asset data (including virtual asset commodity information, related virtual asset transaction data, and virtual asset operation logs, etc.). Classifying these virtual asset data can help people better manage and effectively improve the use efficiency of virtual assets. [0003] At present, my country has carried out research on eID-based virtual asset management and preservation technology in cyberspace to achieve standardized and unified manageme...

AI Technical Summary

Problems solved by technology

In addition, most of the existing algorithms are based on the assumption that the number of samples between categories is basically equal, but in many data, including virtual asset data, the imbalance between categories (labels) is widespread

In addition, some existing studies only use a single or a small number of classifiers for multi-label data training, and the generalization performance is affected to a certain extent.

Images