Eugene Fedorov

By Eugene Fedorov, Tetyana Utkina and Tetyana Neskorodeva

The preliminary stage of the personality biometric identification on a voice is voice signal filtering. For biometric identification are considered and in number investigated the following methods of noise suppression in a voice signal. The smoothing adaptive linear time filtering (algorithm of the minimum root mean square error, an algorithm of recursive least squares, an algorithm of Kalman filtering, a Lee algorithm), the smoothing adaptive linear frequency filtering (the generalized method, the MLEE (maximum likelihood envelope estimation) method, a wavelet analysis with threshold processing (universal threshold, SURE (Stein’s Unbiased Risk Estimator)-threshold, minimax threshold, FDR (False Discovery Rate)-threshold, Bayesian threshold were used), the smoothing non-adaptive linear time filtering (the arithmetic mean filter, the normalized Gauss’s filter, the normalized binomial filter), the smoothing nonlinear filtering (geometric mean filter, the harmonic mean filter, the contraharmonic filter, the α-trimmed mean filter, the median filter, the rank filter, the midpoint filter, the conservative filter, the morphological filter). Results of a numerical research of denoising methods for voice signals people from the TIMIT (Texas Instruments and Massachusetts Institute of Technology) database which were noise an additive Gaussian noise and multiplicative Gaussian noise were received.

Part of the book: Recent Advances in Biometrics

By Eugene Fedorov, Tetyana Utkina and Tetiana Neskorodieva

The preliminary stage of the biometric identification is speech signal structuring and extracting features. For calculation of the fundamental tone are considered and in number investigated the following methods – autocorrelation function (ACF) method, average magnitude difference function (AMDF) method, simplified inverse filter transformation (SIFT) method, method on a basis a wavelet analysis, method based on the cepstral analysis, harmonic product spectrum (HPS) method. For speech signal extracting features are considered and in number investigated the following methods – the digital bandpass filters bank; spectral analysis; homomorphic processing; linear predictive coding. This methods make it possible to extract linear prediction coefficients (LPC), reflection coefficients (RC), linear prediction cepstral coefficients (LPCC), log area ratio (LAR) coefficients, mel-frequency cepstral coefficients (MFCC), barkfrequency cepstral coefficients (BFCC), perceptual linear prediction coefficients (PLPC), perceptual reflection coefficients (PRC), perceptual linear prediction cepstral coefficients (PLPCC), perceptual log area ratio (PLAR) coefficients, reconsidered perceptual linear prediction coefficients (RPLPC), reconsidered perceptual reflection coefficients (RPRC), reconsidered perceptual linear prediction cepstral coefficients (RPLPCC), reconsidered perceptual log area ratio (RPLAR) coefficients. The largest probability of identification (equal 0.98) and the smallest number of coefficients (4 coefficients) are provided by coding of a vocal of the speech sound from the TIMIT based on PRC.

Part of the book: Computational Semantics

By Eugene Fedorov and Tetyana Utkina

An important function of computer systems for automated natural language processing is the analysis of sentiments in text data. Sentiment analysis recognizes text as neutral, negative, or positive, and may also classify more nuanced sentiments, such as angry, happy, sad, or disappointed. It is used to classify big text volumes, such as: chat transcripts and emails; customer reviews; responses of survey; social media posts. The sentiment analysis goal is to identify the emotional tone in text to and implement business decisions, monitor brand reputation, understand sentiment of customer. Types of sentiment analysis is analysis of subtle sentiments; sentiment analysis on based aspect; sentiment detection; sentiment analysis on based intent; sentiment analysis of multilingual. Advantages of sentiment analysis is management of brand reputation; competitor analysis; analysis of customer feedback; innovation quality assessment; marketing campaign evaluation; real-time analysis; improved customer service quality; improved decision-making; analysis scalability; cost-effectiveness. Sentiment analysis includes stages such as text preprocessing, feature extraction from the text, and sentiment classification. The sentiment classification stage is of the greatest interest. At this stage, a shallow machine learning-based approach (performs classification based on metrics, posterior probability, decision trees (ID3 method, C4.5 method, CART method), support vector machines (SVM), and an ensemble of these methods), a deep machine learning-based approach (performs classification based on deep artificial neural networks such as a simple recurrent network (SRN), and one-dimensional convolutional neural networks (CNN)), and a hybrid approach (combining rule-based methods with shallow/deep machine learning methods). The highest classification probability (0.98) is achieved by CNN.

Part of the book: Applied Sentiment Analysis - AI and Machine Learning for Real-World Insights [Working title]