Use of Machine Learning Algorithms for Diagnosis and Treatment of Psychological Pathologies

• Published in: Conference proceedings (IEEE Colombian Conference on Communications and Computing. Online) pp. 1 - 7
• Main Authors: Valverde-Landivar, Galo, Lajones, Karen Charcopa, Rosemberg, Freddy Farino
• Format: Conference Proceeding
• Language: English
• Published: IEEE 21.08.2024
• Subjects: Computational modeling; Computer architecture; deep learning; Emotion recognition; Pathology; Predictive models; Psychology; Spectrogram; speech emotion recognition; Speech processing; Speech recognition; speech spectrograms; Training
• ISSN: 2771-568X
• DOI: 10.1109/COLCOM62950.2024.10720294

Artificial intelligence (AI) and deep learning (ML) have used for training and processing of massive data, allowing the improvement of systems, and making them more intelligent when making decisions. Speech Emotion Recognition (SER) is an area of voice research for speech emotion recognition, evaluating the voice signal and classifying different emotions. In recent years, technological advances in deep learning have helped (SER) to detect and classify emotions effectively, as speech; signal processing methods are difficult due to the variety of emotion frequencies such as happy, angry, sad, neutral and others. In this study, we have used a deep convolutional network architecture (DSCNN) to implement the (SER) model. This uses simple networks to learn salient and discriminative features from the spectrogram of speech signals, generated through the RAVDESS dataset, 8 emotions considered for the analysis and classification of emotions, a prediction result of 61% obtained. Subsequently, an implementation of the (DSCNN) proposed in psychology to determine the diagnoses and treatments of people suffering from depression and anxiety. With the help of this deep neural network, an effective diagnosis obtained in the future and treatment time could reduce.