Improved minimum squared error method for robust classification

In this paper, we improved the method of minimum squared error for robust classification by altering its classification rule. The minimum squared error method, is one of the methods to minimize the sum of the squared error between the output of the linear function and the desired output, which first...

Full description

Saved in:
Bibliographic Details
Main Authors: Fangzhi Zhu, Rui Yan, Yong Sun
Format: Conference Proceeding
Language:English
Subjects:
Algorithm design and analysis
Classification algorithms
Face recognition
Image recognition
Image resolution
Minimum squared error
Pattern recognition
Robustness
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, we improved the method of minimum squared error for robust classification by altering its classification rule. The minimum squared error method, is one of the methods to minimize the sum of the squared error between the output of the linear function and the desired output, which first obtains the mapping that can best transform the training sample into its class label and then exploits the obtained mapping to predict the class label of the test sample. However, the improved method classifies all the test examples by comparing the difference of predicted class labels of test examples and training examples. Through experiment in section 4, we draw a conclusion that the improved MSE (IMSE) method can reduce the classification error and get a better classification result.
ISSN:2376-5933
2376-595X
DOI:10.1109/CCIS.2014.7175705