The object of the study is the process of recognizing the emotional state of students of the distance learning system.
The purpose of the study is to solve the scientific and applied problem related to the recognition of the emotional state of the listeners of the distance learning system by developing an appropriate methodology based on neural network methods and models for the analysis of biometric parameters and aimed at creating appropriate tools that expand the functional capabilities of the distance learning system teaching.
Research methods based on the methodological basis of the theory of complex systems, information theory, pattern recognition and system analysis, the theory of neural networks, wavelet transforms, coding, decision-making, modeling, methods of digital signal processing and biometrics, simulation modeling methods, were used. object-oriented design and elements of the theory of algorithms, programming and experiment planning.
Scientific novelty of research results:
For the first time, the methodology of automated emotional state recognition was developed, which provides the possibility of creating effective emotional state recognition tools that expand the functionality of the distance learning system.
For the first time, a conceptual model for recognizing the emotional state of listeners of the distance learning system was developed, which provided a formalized description of research directions for the development of appropriate means of automated recognition.
For the first time, a method of forming the input field of a neural network model for analyzing the keyboard handwriting of a distance learning system listener was developed, which provides the possibility of effective recognition of the emotions and personality of a distance learning system listener by keyboard handwriting.
For the first time, a method of forming the input field of a neural network model of voice signal analysis was developed, which provides the possibility of effective recognition of emotions and the personality of the listener of the remote learning system by voice.
The approaches to determining the constructive parameters, learning and minimizing the resource consumption of neural network models have been improved, which provided the opportunity to develop effective methods of forming the input field of a neural network model and effective methods of neural network recognition of emotions and the personality of a listener of a distance learning system.
The method of forming the input field of the neural network model of the analysis of biometric parameters associated with images has been improved, which provides the possibility of effective recognition of emotions and the personality of the listener of the distance learning system based on biometric parameters associated with images.
The method of developing a convolutional neural network architecture designed for the analysis of biometric parameters has been improved, which allows to reduce the amount of experimental research related to the development of a convolutional neural network architecture.
The method of neural network recognition of emotions has been improved, which makes it possible to increase the accuracy of recognition of emotions of a listener of a distance learning system based on biometric parameters registered with the help of widespread hardware and software.
We received further development of the model for processing biometric parameters associated with images, which provided a theoretical basis for the development of a method for forming the input field of neural network models for the analysis of biometric parameters associated with images.
The method of using wavelet transforms for filtering the input field of the neural network model for the analysis of biometric parameters associated with images has received further development, which provides the possibility of filtering interference typical for the conditions of the distance learning system.
The method of forming training examples for the neural network model of biometric parameters analysis has received further development, which provides the possibility of reducing the resource intensity of the training process.
The obtained results are related to the implementation of 4 state-budget research projects and are implemented in the educational process of the Kyiv National University of Construction and Architecture, KPI named after Igor Sikorskyi and the National Aviation University.
Recommendations for use: in distance learning systems and in the contours of monitoring the psycho-emotional state of operators of information systems for various purposes; when teaching disciplines related to neural network analysis of biometric parameters.
The obtained scientific results are a methodological basis for the development and implementation of effective tools that have sufficient recognition accuracy and are adapted for use in the conditions of the distance learning system.
