Distance-based features in pattern classification

Tsai, Chih‐Fong; Lin, Wei-Yang; Hong, Zhen-Fu; Hsieh, Chung-Yang

doi:10.1186/1687-6180-2011-62

Cited by 13 publications

(6 citation statements)

References 18 publications

(17 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, other current trend not explored in this paper is the use of unsupervised learning to compute new features from existing ones. Some examples of this are the approach proposed by [40] for intrusion detection or the more general method presented in [41], that expand the original feature vectors by computing new features using distances from each data sample to a number of centroids found by a clustering algorithm. In this same line, the use of distance-based features to a set of reference patterns, or the related concept of pairwise dissimilarities [42], could effectively be used to enrich or replace the information provided in the original feature vectors.…”

Section: Discussionmentioning

confidence: 99%

Fusing highly dimensional energy and connectivity features to identify affective states from EEG signals

2017

View full text Add to dashboard Cite

In this paper, a novel method for affect detection is presented. The method combines both connectivity-based and channel-based features with a selection method that considerably reduces the dimensionality of the data and allows for an efficient classification. In particular, the Relative Energy (RE) and its logarithm in the spacial domain, and the spectral power (SP) in the frequency domain are computed for the four typical frequency bands (α, β, γ and θ), and complemented with the mutual information measured over all channel pairs. The resulting features are then reduced by using a hybrid method that combines supervised and unsupervised feature selection. First, a Welch's t-test is used to select the features that best separate the classes, and discard the ones that are less useful for classification. To this end, all features where the t-test yields a p-value above a threshold are eliminated. The remaining ones are further reduced by using Principal Component Analysis. Detection results are compared to state-of-the-art methods on DEAP, a database for emotion analysis composed of labeled recordings from 32 subjects while watching 40 music videos. The effect of using different classifiers is also evaluated, and a significant improvement is observed in all cases.

show abstract

Section: Discussionmentioning

confidence: 99%

Fusing highly dimensional energy and connectivity features to identify affective states from EEG signals

2017

View full text Add to dashboard Cite

show abstract

“…The k-means algorithm is generally used to identify centroids (averaging object) in many studies on distance-based feature extraction [7], [43] since it is easy to realize and its computational cost is low. However, k-means has two drawbacks.…”

Section: ) Energy Normalizationmentioning

confidence: 99%

“…After extracting the wavelet ridge, we employ three preprocessing methods on the wavelet ridge: the elimination of the influence of noise using SVD denoising, which has a time complexity of (n 2 ), dimensionality reduction using PAA which has a time complexity of (m), where m denotes the reduced dimension, and normalization with respect to energy, which has a linear time complexity of (m). For the existing M -dimensional training data, the time complexity of the DBA algorithm is (I • M • m 2 ) [43], where the parameter I denotes the number of iterations. Obviously, when large training datasets or long signal lengths appear, time complexity will increase rapidly.…”

Section: E Computational Complexity Analysismentioning

confidence: 99%

Automatic Modulation Recognition of Radar Signals Based on Manhattan Distance-Based Features

Huang

Jin

et al. 2019

IEEE Access

View full text Add to dashboard Cite

Feature-based (FB) algorithms for automatic modulation recognition of radar signals have received much attention since they are usually simple to realize. However, existing FB approaches usually focus on several specific modulations and fail when applied to various modulations. To overcome this issue, we propose a simple and effective FB algorithm based on Manhattan distance-based features (MDBFs) in this paper. MDBFs are new features for radar signals that can be applied for recognition of different modulations. The main contributions of this paper are as follows. First, radar signals are represented as wavelet ridges, which includes important information that can distinguish different modulations, and the piecewise aggregate approximation algorithm is introduced to reduce signal dimensions. Then, the dynamic time warping averaging is employed instead of the traditional k-means algorithm to extract realistic centroids for each class. Finally, the Manhattan distances between each data sample and each centroid are used to construct MDBFs, and decisions are made using the k-nearest neighbor. In addition, we prove that MDBFs have better class separability power than the Euclidean-based features. MDBFs contain information about the correlations between different classes, which means that these features suitable for discriminating various modulations when their class distributions do not overlap badly in representation space. The extensive experiments on a synthetic dataset demonstrate the outstanding performance of our proposed method and are hardly affected by the pulse width of the signal. Thus, the proposed method with the effectiveness and robustness could be a promising modulation recognition method of the radar signal. INDEX TERMS Modulation recognition, Manhattan distance-based feature, wavelet ridge, dynamic time warping averaging, class centroid.

show abstract

“…For this purpose, the multicategory support vector machine (M-SVM) method can be used. For more details in this regard, one can refer to [13,22].…”

Section: Estimation Of Subsystems and Subregionsmentioning

confidence: 99%

“…Also in [21], a method using selforganizing map (SOM)-based spectral clustering is proposed for agriculture management. In [22], a novel distance-based feature extraction method for various pattern classification problems is introduced. Specifically, two distances are extracted, which are based on (1) the distance between the data and its intra-cluster center and (2) the distance between the data and its extracluster centers.…”

Section: Introductionmentioning

confidence: 99%

Identification of piecewise affine systems based on fuzzy PCA-guided robust clustering technique

Khanmirza

Nazarahari

Mousavi

2016

EURASIP J. Adv. Signal Process.

View full text Add to dashboard Cite

Hybrid systems are a class of dynamical systems whose behaviors are based on the interaction between discrete and continuous dynamical behaviors. Since a general method for the analysis of hybrid systems is not available, some researchers have focused on specific types of hybrid systems. Piecewise affine (PWA) systems are one of the subsets of hybrid systems. The identification of PWA systems includes the estimation of the parameters of affine subsystems and the coefficients of the hyperplanes defining the partition of the state-input domain. In this paper, we have proposed a PWA identification approach based on a modified clustering technique. By using a fuzzy PCAguided robust k-means clustering algorithm along with neighborhood outlier detection, the two main drawbacks of the well-known clustering algorithms, i.e., the poor initialization and the presence of outliers, are eliminated. Furthermore, this modified clustering technique enables us to determine the number of subsystems without any prior knowledge about system. In addition, applying the structure of the state-input domain, that is, considering the time sequence of input-output pairs, provides a more efficient clustering algorithm, which is the other novelty of this work. Finally, the proposed algorithm has been evaluated by parameter identification of an IGV servo actuator. Simulation together with experiment analysis has proved the effectiveness of the proposed method.

show abstract

Distance-based features in pattern classification

Cited by 13 publications

References 18 publications

Fusing highly dimensional energy and connectivity features to identify affective states from EEG signals

Fusing highly dimensional energy and connectivity features to identify affective states from EEG signals

Automatic Modulation Recognition of Radar Signals Based on Manhattan Distance-Based Features

Identification of piecewise affine systems based on fuzzy PCA-guided robust clustering technique

Contact Info

Product

Resources

About