Performance of Combined Support Vector Machine and Principal Component Analysis in recognizing infant cry with asphyxia

Abstract

Combined Support Vector Machine (SVM) and Principal Component Analysis (PCA) was used to recognize the infant cries with asphyxia. SVM classifier based on features selected by the PCA was trained to differentiate between pathological and healthy cries. The PCA was applied to reduce dimensionality of the vectors that serve as inputs to the SVM. The performance of the SVM utilizing linear and RBF kernel was examined. Experimental results showed that SVM with RBF kernel yields good performance. The classification accuracy in classifying infant cry with asphyxia using the SVM-PCA is 95.86%.

Quantitative assessment of cancer cell morphology and motility using telecentric digital holographic microscopy and machine learning

“…A linear support vector machine learning algorithm was used based on excellent classification performance in previous studies using DHM data sets . Further, the method of PCA to reduce the dimensionality of features extracted for machine learning is well‐established in the literature . The use of principal components reduced co‐linearity of features used as predictors for machine learning, and avoided overfitting error.…”

Classifying infant cry patterns by the Genetic Selection of a Fuzzy Model

“…Please note that even though other works have tackled the infant cry classification problem, a direct comparison against those methods was not performed; this is because they do not apply 10-fold cross validation. Nonetheless, accuracy percentages of 99% and of 95% are reported in [26] and [48], respectively, for asphyxia vs. normal cries; this values can be used as reference, although we emphasize that results are not comparable at all.…”

“…Lederman et al [33] Healthy infants and no healthy 20-MFCC Hidden Markov Models 63.00% Orozco-Garcia and Reyes-Garcia [41] Normal and pathological cries LPC Neural Networks 94.30% Cano Ortiz et al [13] Normal and abnormal Statistical Neural Networks 85.00% Reyes-Galaviz et al [44] Normal, deaf, and asphyxia MFCC ANFIS 96.67% Barajas and Reyes [5] Normal, deaf, and asphyxia MFCC/LPC FRNN 91.08% (overall) Lederman et al [34] Clef palate 20-MFCC Hidden Markov Model 91% Sahak et al [48] Normal and asphyxia MFCC SVM 95.86% Hariharan et al [26] Normal and pathological Wavelet PNN 99.00%…”

“…Among these techniques, artificial neural networks have been one of the most widely used [13,26,41,45]. With the same purpose, the use of support vector machines (SVMs) [7,48], hidden Markov models [33,34], as well as several hybrid approaches that combine fuzzy logic with neural networks [44,[50][51][52], fuzzy logic with support vector machines [6] or evolutionary strategies with neural networks [23] have also been explored. Table 1 summarizes the characteristics of previous studies around the infant cry recognition.…”

Comparison of tasselled cap transformations based on the selective bands of Landsat 8 OLI TOA reflectance images

“…Therefore, feature extraction and image classification algorithms based on PCA and SVM were developed. It has been proved that such algorithms could reduce computational time with the reduced dimensionality of features, while improving classification accuracy (Tan, Lani, and Lai 2007;Sahak et al 2010). Compared with PCA components, TCTB, TCTG, and TCTW have specific physical characteristics and capture the majority of variations associated with land cover information, while reducing the dimensionality of the spectral bands.…”