Generalized MCE Training Algorithm for Feature Dimensionality Reduction
نویسندگان
چکیده
Dimensionality reduction is an important problem in pattern recognition. Reducing the dimensionality of feature can improve the effecitveness and efficiency of pattern recognition algorithms. Minimum Classification Error(MCE) training algorithm is a power tool for dimensionality resuction. However, MCE training process is a type of thorough search process for the local minimum, global minimum can not be guaranteed by this process. In this paper, a generalized MCE training algorithm is proposed. This algorithm uses a general search process for searching the generalized starting point. Then conventional MCE training algorithm is used to search for the minimum.
منابع مشابه
Feature extraction and dimensionality reduction algorithms and their applications in vowel recognition
Feature extraction is an important component of a pattern recognition system. It performs two tasks: transforming input parameter vector into a feature vector and/or reducing its dimensionality. A well-de3ned feature extraction algorithm makes the classi3cation process more e4ective and e5cient. Two popular methods for feature extraction are linear discriminant analysis (LDA) and principal comp...
متن کاملA Modified Minimum Classification Error (MCE) Training Algorithm for Dimensionality Reduction
Dimensionality reduction is an important problem in pattern recognition. There is a tendency of using more and more features to improve the performance of classifiers. However, not all the newly added features are helpful to classification. Therefore it is necessary to reduce the dimensionality of feature space for effective and efficient pattern recognition. Two popular methods for dimensional...
متن کاملUsing Minimum Classification Error Training in Dimensionality Reduction
Dimensionality reduction is an important problem in pattern recognition. I n a speech recognition system, the size of the feature set is normally large in the order of 40. Therefore, it is necessary to reduce the dimensionality of the feature space for efficient and effective speech recognition. Two popular methods to reduce the dimensionality of the feature space are Linear Discriminat Analysi...
متن کاملReview of Minimum Classification Error Training in Dimensionality Reduction
Several modelling techniques are used in speech recognition to model the short term variations in a speech signal. These techniques generally use a high dimensional feature vector which could be correlated. Several classical techniques of discriminant analysis are used for reducing the dimensionality of the input feature without affecting the overall performance. One such approach is the minimu...
متن کاملApplying Genetic Algorithm to EEG Signals for Feature Reduction in Mental Task Classification
Brain-Computer interface systems are a new mode of communication which provides a new path between brain and its surrounding by processing EEG signals measured in different mental states. Therefore, choosing suitable features is demanded for a good BCI communication. In this regard, one of the points to be considered is feature vector dimensionality. We present a method of feature reduction us...
متن کامل