Active learning classification method based on Gaussian mixture model and sparse Bayesian

Abstract

The invention discloses an active learning classification method based on a Gaussian mixture model and sparse Bayesian, and the method comprises the following steps of 1) employing the Gaussian mixture model to train all samples including marked samples and unmarked samples, and obtaining the mixing coefficient, the mean value, and the covariance of each Gaussian component; 2) constructing an initial training set XL, taking the initial training set as a marked sample set, and updating an unmarked sample set XU; 3) constructing an initial GMM-FRVM (Gaussian mixture kernel-based direct-push typerelevance vector machine) model; 4) updating the marked sample set and the unmarked sample set based on the GMM-FRVM model; 5) retraining the GMM-FRVM model based on the updated marked sample set andthe updated unmarked sample set; and 6) finishing the classification marking of all samples by adopting the final GMM-FRVM model. According to the method disclosed by the invention, a better classification effect is obtained through artificial marking as few as possible by means of the active learning combining the Gaussian mixture model and the sparse Bayesian.

Description

technical field [0001] The invention relates to the field of machine learning, in particular to an active learning classification method based on a mixed Gaussian model and sparse Bayesian. Background technique [0002] With the rapid development of computer-related technologies, society has become more information-based, and a large amount of data is generated every day. In actual scenarios, a large amount of data obtained by people is unlabeled. When traditional supervised learning methods use a small amount of labeled data for training, it is difficult to achieve better prediction results due to small sample size and less information. Manually labeling samples takes a lot of time and energy, and in some cases it is even impossible to complete the labeling of a large number of samples. Based on these problems, the present invention proposes an active learning classification method based on mixed Gaussian model and sparse Bayesian. [0003] Active learning expands the sam...

AI Technical Summary

Problems solved by technology

As a supervised learning method, the traditional correlation vector machine only uses the labeled data as the training set to construct the learning model, which is easy to cause problems such as small sample size and poor predictive performance.

Images