Speech Emotion Recognition (SER) is a very interesting task that allows the machine to identify and recognize different emotional states from human speech using new technologies. The SER can be represented by two main steps, namely feature extraction emotion classification. Our contribution field will focus on these phases. This paper seeks investigate influence of embedded features in Wav2vec2 HuBERT models SER, variants per module are implemented, including base, large, large X-large. In...

Because one of the key issues in improving performance Speech Emotion Recognition (SER) systems is choice an effective feature representation, most research has focused on developing a level fusion using large set features. In our study, we propose relatively low-dimensional that combines three features: baseline Mel Frequency Cepstral Coefficients (MFCCs), MFCCs derived from Discrete Wavelet Transform (DWT) sub-band coefficients are denoted as DMFCC, and pitch based Moreover, proposed...

Crude oil prices play a critical role in the global economy, and accurately predicting their future values is of great importance. In this study, we evaluated performance three support vector machine (SVM) models - standard SVM, SMO-based SGD-based SVM crude based on daily, weekly, monthly data, aligning with Industry 4.0 paradigm, to enhance decision-making gas sector. The results showed that all performed well, high coefficient determination R2 low MSE across versions dataset. SMO...

In the evolving landscape of global energy, accurately forecasting oil prices plays a pivotal role in strategizing transition to green energy. Traditional methods, though widely used, often fall short capturing multifaceted influences on price dynamics. This research introduces multimodal deep learning approach that incorporates both time series data and key economic indicators enhance accuracy. By employing combination Long Short-Term Memory (LSTM) Temporal Convolutional Networks (TCN),...

Feature Selection (FS) is one of the power solutions used in Machine Learning (ML) problems, since it can help to remove irrelevant and redundant attributes, improve performance, reduce computation time build more robust models. In this work, a thorough study carried out examine effect well-performing filter embedded FS methods for credit scoring. Further, we explore such on prediction models obtained using different classification techniques. We conduct our experiments Australian dataset...

Speech Emotion Recognition (SER) refers to the ability of Machine Learning (ML) and Deep (DL) techniques accurately predict people's emotional states from speech signals. significant progress has been achieved in SER domain involving incorporation DL models introduce novel features extraction processes. This paper introduces use deep representations learned multi-modal Large Language Model (LLM) called ImageBind. These were subsequently provided as input Nu-Support Vector (Nu-SVM) with RBF...

This paper examines the development of a speaker identification system (SIS) for future aeronautical communication systems. SIS promises to improve flight safety by reducing incidence call-sign confusion events. However, practical such faces many challenges, especially related signal corruption channel noise. Due dynamic motion aircraft, experiences high Doppler shifts and fading due multipath propagation. means that is required be robust against perturbations. In proposed system, noise was...

Effective Speaker Identification System (SIS) involves extracting features effectively. In this paper, we propose a feature extraction scheme based on wavelet analysis which is used along with short-term features. To overcome the drawbacks of Discrete Wavelet Transform (DWT), to combine Stationary (SWT) Mel-Frequency Cepstral Coefficient (MFCC) The combined were as inputs K-nearest neighbors (Knn) classifier. effectiveness proposed method investigated for closed-set text-independent SIS in...