Imbalance sample weighting method suitable for training of support vector machine

Abstract

The invention is mainly used in the field of artificial intelligence and relates to an imbalance sample weighting method suitable for training of a support vector machine. The imbalance sample weighting method comprises the steps of utilizing clustering and Fisher discriminant rate criterion to perform redundant data reduction, then calculating the distances between reduced data samples and a fuzzy classification face, endowing corresponding weights according to the distances and then using the weighted data samples to perform training of the support vector machine. For solving the problem that a traditional support vector machine is still needed to be modified and improved on the aspect of large dataset processing or imbalance data sample processing, a novel algorithm is provided, corresponding weighting is conducted on the reduced large sample data again so as to perform training learning of the support vector machine, the training speed of the support vector machine is improved, the classification accuracy of the support vector machine is also improved, and great benefit is brought to classification of datasets of large samples.

Description

technical field [0001] The present invention is mainly used in the field of artificial intelligence, especially the technology of pattern recognition, relates to redundant data reduction based on clustering and Fisher's discriminant rate and a method for weighting unbalanced samples, especially a method suitable for supporting Imbalanced sample weighting method for vector machine training. Background technique [0002] Data classification has always been an important application branch in the field of artificial intelligence such as pattern recognition, and is widely used in character recognition, face detection and recognition, etc. There are a variety of classification techniques currently available, including decision tree methods, neural network methods, and support vector machine methods. The support vector machine method has gradually developed into the most widely used and most prominent classification effect because of its scientific statistical learning theoretical ...

AI Technical Summary

Problems solved by technology

[0003] Although the above-mentioned algorithms have made great improvements on the traditional support vector machine method, and have shown their respective advantages in dealing with large data and unbalanced data, they have ignored the fact that only those near the final classification decision surface in the training data set are Only the outer data samples may become support vectors, and most of the remaining samples play a small or no role in training the support vector machine. It is a space and time to weight these redundant data whose practical significance can be ignored. waste on

Images