Analyzing social psychological impact on emotional expression through peer communication using crayfish optimization algorithm with deep learning model

In today’s digital age, people frequently interact with multiple devices simultaneously, significantly reshaping how they express emotions and communicate with peers. The insights gained will advance the fields of social psychology and human-computer interaction (HCI), informing the design of digital platforms that better support meaningful emotional and social interactions. Sentiment analysis (SA) identifies people’s emotions, attitudes, and sentiments towards a given target, like activities, people, services, organizations, products, and subjects. Emotion detection is a subdivision of SA as it forecasts the novel emotion instead of only maintaining negative, positive, or neutral. Emotion recognition has emerged as an important area of study that may report different valuable inputs. Emotion is expressed in numerous ways that are observed, namely written text, gestures, speech, and facial expressions. Emotional recognition in the text document is primarily a content-based classification problem containing ideas from natural language processing (NLP). NLP methods enhance the performance of learning-based models by combining the syntactic and semantic features of the text. To identify the emotion, a new deep learning (DL) model is applied to recognize emotional expression from text for improved results. This paper uses the Crayfish Optimization Algorithm and Deep Learning (SPIEEPC-COADL) method to analyze the Social Psychological Impact on Emotional Expression through Peer Communication. The presented SPIEEPC-COADL model aims to develop an effective method for detecting text-based emotional expressions to enhance HCI. Initially, the text pre-processing stage contains various levels to clean, normalize, and structure raw text data to improve the performance. Furthermore, the FastText method is employed for the word embedding process. Moreover, the variational autoencoder (VAE) model is implemented for emotion classification. Finally, the crayfish optimization algorithm (COA) adjusts the VAE model’s hyperparameter values, improving classification. The efficiency of the SPIEEPC-COADL model is examined using emotion detection from the text dataset. The comparison study of the SPIEEPC-COADL technique demonstrated a superior accuracy value of 99.07% over existing models.