A fuzzy K-nearest neighbor classification method and system based on weighted chi-square distance metric

Abstract

The invention discloses a fuzzy K-nearest neighbor classification method and system based on weighted chi-square distance metric, belonging to the field of fuzzy K-nearest neighbor algorithm. The technical problem to be solved is that the classification accuracy of the existing fuzzy K-nearest neighbor algorithm is low, and the classification accuracy is influenced by noise, irrelevant data and unweighted data. The method comprises the following steps: establishing a training sample and a test sample, and setting a value of a nearest neighbor number k; Calculating weighted chi-square distance;with weighted chi-square distance between the training sample and the test sample as the distance measure, by using the class membership degree of k nearest neighbor training samples,determining theclass of the test sample; evaluating the class. The system includes a sample building module, a nearest neighbor number setting module, a weighted chi-square distance calculation module, a sample category determination module and a sample evaluation module. The classification time is short, the classification accuracy is not affected by noise, relative data and unweighted data, and the stability of the method is high.

Description

technical field [0001] The present invention relates to the field of fuzzy K-nearest neighbor algorithms, in particular to a fuzzy K-nearest neighbor classification method and system based on a weighted chi-square distance measure. Background technique [0002] In the classification method, KNN is an inert classification learning algorithm, that is, the learning model is not constructed until a test sample is given. It classifies according to the class of samples that are the majority among the K nearest neighbor samples to the test sample. The specific process is as follows: Assume that each category contains multiple samples, and each sample belongs to a specific category. KNN first calculates the distance between each known sample and the sample to be divided, and then takes the category that is the majority of the K samples closest to the sample to be divided as the category of the sample to be divided. Ideally, a distance metric must minimize the distance between two ...

AI Technical Summary

Problems solved by technology

[0006] The technical task of the present invention is to address the above deficiencies, providing a fuzzy K-nearest neighbor classification method and system based on the weighted chi-square distance measure, to solve the problem that the existing fuzzy K-nearest neighbor algorithm has a low classification accuracy rate, and the classification accuracy rate is affected by noise and irrelevance. Data and issues with unweighted data impact

Images