Central AsianResearch Index

AUTOMATING PUBLIC SPEAKING ANXIETY TREATMENT USING LSTM RECURRENT NEURAL NETWORK AND AUGMENTED REALITY

Public Speaking Anxiety (PSA) is recognized as one of the most prevalent fears, significantly impacting individuals in both professional and daily contexts. While traditional psychological treatments have been enhanced by technologies like Virtual Reality (VR) and Augmented Reality (AR), less research has focused on automating the treatment model to reduce therapy time. Therefore, the aim of this research is to propose a novel model to automate the treatment process for PSA using an adaptive LSTM recurrent neural network. To train the proposed model, a dataset consisting of physiologicalinformationfrompatientstreatedbyapsychologistwasused. Considering theassignmentofdifferentweightstothetrainingdatabythepsychologistbasedonthe patient’s improvement level, a weighted loss function (Weighted MSE) is proposed for training the suggested LSTM neural network model. The comparative results between the weighted MSE trained by LSTM and the unweighted MSE demonstrated that the LSTMalgorithmwiththeweightedMSEimprovesdiagnosticperformance. Thismeans that the proposed model can facilitate and accelerate patient treatment in accordance with the psychologist’s opinion, which justifies and even recommends its use.

Сведения

Опубликовано
2026 · Т. 14, № 2, С. 4-34
Язык
Английский
ISSN
2306-6172 · 2308-9822
Издатель
L. N. Gumilyov Eurasian National University
Аффилиации
Islamic Azad University
Payame Noor University