Emotion Classification for Hindi Text: A Hybrid Approach

In recent years, with the popularity of the Internet, more and more people like to comment on movies they have watched on the film platform after watching them. These reviews hide the reviewers' feedback on films. Mining the emotional orientation information in these reviews can provide consumers with shopping references and help businesses optimize film works and improve business strategies. Therefore, the emotional classification of film reviews has high research value because few emotion dictionaries and analysis tools are available for reference and use in film reviews. The accuracy of emotion classification still needs to be improved. This study introduces the attention mechanism and dual channel long short term memory (DC-LSTM) while building the emotion dictionary in the field of Chinese film review. It classifies Chinese film reviews in terms of topic-based fine-grained emotion. First, the emotion vector is constructed using the constructed movie review emotion lexicon. The semantic vector obtained by the Word2vector tool is input to LSTM to encode the comment text. Then, the topic attention module is used to decode. Finally, the final emotion classification result is obtained through the softmax function of the entire link layer and the output layer. The thematic attention modules constructed in this study are independent of each other for attention parameter adjustment and learning. One attention module corresponds to one film theme. In this study, eight themes, including "plot," "special effects," "original work," "music," "thought," "theme," "acting skills," and "joke," were extracted, and each theme was classified into three types of emotions: "positive," "neutral," and "negative." The experimental results on the crawled Chinese film review dataset show that the proposed algorithm is superior to some existing algorithms and models in accuracy, precision, recall and F1 measure. The DCLSTM based on the thematic attention mechanism (DCLSTM-TAM) model constructed in this study introduces the emotion vector into the network and adds the theme attention mechanism. It can not only classify the emotion for different topics of a film review but also effectively deal with film reviews with fuzzy emotional tendencies. It realizes the fine-grained emotion classification of film topics and improves the accuracy of emotion classification of film reviews. The emotion classification method and model proposed in this study have good transferability, and the change of training corpus is also applicable to other short text fields.

Emotion analysis has emerged as an important branch of Natural Language Processing (NLP) for identifying and classifying human emotions expressed in textual data. In the Indonesian context, emotion analysis plays a significant role in understanding public perceptions and improving government services. This study focuses on emotion classification in public service complaints submitted to the Dinas Perhubungan Kabupaten Jember. The dataset consists of 259 comments collected through a questionnaire, each labeled with one of three emotions: senang (joy), kecewa (sad), and netral (neutral). Two approaches were evaluated: TF-IDF + SVM, representing a traditional statistical model, and a Hybrid model (TF-IDF + Word2Vec) + SVM, which integrates semantic word embeddings with statistical features. Experimental results show that TF-IDF + SVM achieved the highest accuracy (0.784), indicating strong performance in classifying dominant complaint categories. However, its relatively lower recall (0.651) suggests limitations in identifying complaints expressed with alternative words. The Hybrid model, while slightly lower in accuracy (0.780), demonstrated improvements in recall (0.655) and F1-score (0.682), reflecting more balanced performance. This improvement highlights the contribution of semantic features in capturing diverse linguistic expressions within complaints. The findings suggest that TF-IDF + SVM is suitable when accuracy is prioritized, while the Hybrid model is preferable for ensuring broader coverage of citizen concerns.

<p><span id="docs-internal-guid-c2be589c-7fff-55df-f7fc-a27d9748d2d4"><span>Human emotion recognition</span><span> is an upcoming research field of human computer interaction based on facial gestures and is being used for real-time analysis in classifying cognitive affective states from a facial video data. Since computers have become an integral part of life, many researchers are using emotion recognition and classification of data based on audio and text. But these approaches offer limited accuracy and relevance in emotion classification. Therefore we have introduced and analyzed a hybrid approach which could outperform the existing strategies that uses an innovative approach supported by selection of audio and video data characteristics for classification. The research uses SVM for classifying the data using audio-visual savee database and the results obtained show maximum classification accuracy with respect to audio data about 91.6 could be improved to 99.2% after the application of hybrid strategy. </span></span></p>

In recent years, the integration of artificial intelligence has revolutionized human-computer interaction by enabling devices to provide personalized user experiences. This study focuses on the development of a Smart Mirror with Emotion Monitoring, which utilizes advanced facial recognition and deep learning algorithms to detect and classify user emotions in real time. By analyzing facial expressions, the system identifies emotional states such as happiness, sadness, surprise, angry, and neutrality. Based on these detected emotions, the mirror responds dynamically by displaying personalized content, including motivational messages, and aiming to enhance mental well-being. Beyond emotional responsiveness, the smart mirror also functions as a daily assistant, offering real-time information such as weather updates, calendar, and emotion monitoring insights. Its applications extend beyond residential use to sectors such as healthcare, retail, and corporate environments, where emotion-based interactions can improve user engagement and experience. In healthcare, for instance, the mirror can assist individuals dealing with stress or mental health conditions by providing therapeutic interactions. Similarly, in commercial settings, the mirror can be integrated into customer service environments to enhance user satisfaction. The proposed system is built using state-of-the-art computer vision techniques, deep learning models, and IoT integration to ensure accurate emotion detection and seamless functionality. A robust dataset is employed to train and optimize the facial recognition model, ensuring high accuracy in emotion classification. To validate the system’s performance, extensive experiments are conducted, evaluating parameters such as detection accuracy, response time, and user satisfaction. The results demonstrate the system’s effectiveness in providing real-time emotional insights and personalized interactions, contributing to the advancement of intelligent assistive technologies. By incorporating emotional intelligence into smart mirror technology, this study bridges the gap between functionality and user-centric design, paving the way for next-generation AI-driven smart devices

Strong emotions can be expressed and evoked via music. However, accurately recognizing the emotions in music using computational models is extremely challenging. When the music portions convey various complex emotions, the problem’s difficulty can rise significantly. In this work, a systematic strategy that blends multiple cutting-edge techniques is used to focus on the emotional classification of music. The Kaggle-sourced music emotion dataset provides a wide range of audio samples illustrating various emotional expressions in music. Spectral subtraction as a noise reduction approach is used to improve the quality of the analysis. This efficiently reduced unwanted background noise and improved the audio signals’ clarity. A Quantum Convolutional Neural Network (QCNN) is used for feature extraction, taking advantage of its special capacity to extract complex patterns and characteristics from the music data that are essential for emotion recognition. To provide accurate predictions, the training procedure is optimized and finally utilized Flexible Runge Kutta Optimized Multilayer Perceptrons (FRKO-MLP) for emotion classification. The goal of this combination of advanced classification approaches, quantum feature extraction and noise reduction is to increase the robustness and accuracy of music emotion identification systems. Based on the chosen datasets, the results of the experimental study demonstrate that the suggested model greatly increases the efficiency and accuracy of emotion classification in music. The proposed FRKO-MLP attains 93.21% accuracy, 90.52% precision, 87.34% sensitivity and 89.72% F1-score. These findings show that the model is far better at identifying emotions than traditional methods.

Emotion plays a very significant role in human's everyday communication. Emotion recognition is an essential activity for a computer to understand the human inner state in the Internet of things (IoT)-based brain–computer interface (BCI) framework. Researchers nowadays are more intent toward emotion detection from electroencephalogram (EEG) signals. In recent years, many researchers have conducted various methods of emotion detection from physiological (EEG) signals. A variety of features are extracted from the EEG signals, and various classification methods are applied to these features. In this chapter, different classifiers based on machine learning such as K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), and Gaussian Naive Bayes (GNB) with Principal Component Analysis (PCA) are used to simulate the relationship between emotional states and EEG signals of the brain. DEAP dataset is used to test the efficacy of these machine learning techniques for the analysis of emotions using physiological EEG signals and it is realized that SVM with PCA significantly improves accuracy in cases of valence, arousal, and dominance. In the case of liking, KNN with PCA gives better results in terms of accuracy. It is also found that SVM with PCA provides better classification accuracy i.e. 64.06% for valence, 64.30% for arousal, 69.50% for dominance, and 64.06% for liking, respectively.

Emotional analysis techniques for text are very useful for analyzing public opinions on social media, and user reputation and opinions about products and services on electronic commerce sites. The Chinese Emotional Word Dictionary (CEWD) was created and has a very rich vocabulary for the fine-grained (11 emotion categories) emotional analysis of Chinese text. Additionally, the Chinese Emotional Expression Analysis System (CEEAS) that uses CEWD was developed to analyze emotions in Chinese text relatively accurately and quickly. In this study, we apply CEEAS to the emotional analysis of modern Chinese literature to investigate the characteristics of the appearance of emotional categories in modern Chinese literary works. Then, we apply CEEAS to modern translations of ancient Chinese literary works to perform emotional analysis, and compare the appearance characteristics of emotional expressions with those of modern literary works. Furthermore, we report on the changes in the appearance characteristics of emotional expressions between modern Japanese novels and their Chinese translations. We are convinced that literary research using the emotional analysis techniques proposed in this study will be useful as a new digital approach to literary research.

Emotion is an essential aspect of human life, and effectively identifying corresponding emotions from different scenarios will help promote the development of human-computer interaction systems. Therefore, emotion classification has gradually become a challenging and popular research field. Compared with text emotion analysis, emotion analysis of audio data is still relatively immature. Traditional audio sentiment analysis research is based on feature information such as MFCC, MFSC, etc. while using time-memory models such as LSTM and RNN for emotion analysis. Due to the rapid development of transformers and attention mechanisms, many scholars have shifted their research from the RNN family to the transformer family or deep learning models with attention mechanisms. Therefore, this paper proposes a method to convert audio data into a spectrogram and use a vision transformer model based on transfer learning for emotion classification. This paper conducts experiments on the IEMOCAP dataset and the MELD dataset. The experimental results show that the emotion classification accuracy of the Vision transformer in the IEMOCAP and the MELD datasets reach 56.18% and 37.1%.

Advancements in neuroscience and deep learning have significantly enhanced bio-signal-based emotion recognition, a critical component in Brain-Machine Interface (BMI) applications for healthcare, human-computer interaction, and human-AI assistant communication. Former studies have proposed Manual Mapping electrode matrices and employing Convolutional Neural Networks (CNNs) to recognize spatial EEG activities. However, this Manual Mapping of EEG electrodes onto matrix grids limits spatial precision and introduces inefficiencies. This study proposes automated channel mapping methods of Orthographic Projection and Stereographic Projection to address these challenges, using Differential Entropy and Power Spectral Density with Linear Dynamical Systems as features. A 3-branch multiscale CNN was trained on open-source dataset, employing a 5-fold cross-classification approach. Experimental results demonstrate that higher-resolution grids (16×16, 24×24) with automated projections significantly outperform Manual Mappings, achieving up to a 4.06% improvement in classification accuracy (p < 0.05). This result indicates that enhancing spatial precision of EEG data improves emotion classification, establishing automated spatial mapping as an advancement in EEG-based emotion recognition.Clinical Relevance-Advancement in emotion classification accuracy can facilitate more reliable diagnostic tools and personalized therapeutic interventions for mental health disorders, such as depression and anxiety.

With a rapid growth of the internet communication, many types of text are produced. They can convey the meanings that can contribute to text categorization. Emotion classification also becomes more interesting, but emotion classification in Thai text is still not able to be correctly classified. Thus, this paper proposes a novel approach that takes advantage of bi-words occurrence to classify emotion hidden in a short sentence. In this paper, we classify Thai text into six basic universal emotions including anger, disgust, fear, happiness, sadness, and surprise based on latent semantic analysis approach. We compared the results between two models which construct features from the sentences and applied to three classification methods, i.e. Naive Bayes, SVM, and Decision Tree. The first feature model uses only single word occurrence in the classification. The second model uses single word combined with bi-words occurrence in the classification. The results show that the second model can yield higher accuracy than the first model based on the Naive Bayes classification method.

This paper presents a real-time facial recognition system tailored for advertising personalization. The proposed platform analyzes facial attributes: age, gender, and emotion from live image or video feeds to dynamically deliver targeted advertisements. The system leverages deep learning models (Convolutional Neural Networks and recurrent networks) to perform robust face analysis and couples these predictions with a content-based recommendation engine. We implement the solution as a web application using Flask, TensorFlow, and OpenCV, integrating pre-trained CNN architectures (ResNet50, VGG16, GoogleNet) and an LSTM for temporal emotion modeling. The system is trained on large-scale face datasets (UTKFace for age/gender, FER2013 for expressions) and achieves high accuracy in demographic and emotion classification, which translates into effective ad recommendations. To address ethical considerations, we incorporate privacy-by-design principles, no personal identifiers are stored, and all processing is done in- memory, and we align with legal standards (e.g. GDPR) to protect user data. Experimental evaluation demonstrates 96% gender classification accuracy, a 4-year mean absolute error in age estimation, and an average F1-score of 0.72 in emotion recognition, enabling a recommendation match F1 of 0.87. These results highlight the potential of deploying deep learning-based facial analytics in commercial advertising for enhanced user engagement, while maintaining responsible AI practices and user trust.

Sentiment analysis of netizens’ comments can accurately grasp the psychology of netizens and reduce the risks brought by online public opinion. However, there is currently no effective method to solve the problems of short text, open word range, and sometimes reversed word order in comments. To better solve the above problems, this article proposes a hybrid model of sentiment classification, which is based on bidirectional encoder representations from transformers (BERT), bidirectional long short-term memory (BiLSTM) and a text convolution neural network (TextCNN) (BERT-BiLSTM-TextCNN). The experimental results show that (1) the hybrid model proposed in this article can better combine the advantages of BiLSTM and TextCNN; it not only captures local correlation while retaining context information but also has high accuracy and stability. (2) The BERT-BiLSTM-TextCNN model can extract important emotional information more flexibly in text and achieve multiclass classification tasks of emotions more accurately. The innovations of this study are as follows: (1) the use of BERT to generate word vectors has the advantages of more prior information and a full combination of contextual semantics; (2) the BiLSTM model, as a bidirectional context mechanism model, can obtain contextual information well; and (3) the TextCNN model can obtain important features well in the problem of text classification, and the combined effect of the three modules can significantly improve the accuracy of emotional multilabel classification.

Focusing on emotion recognition, this paper addresses the task of emotion classification and its performance with respect to accuracy, by investigating the capabilities of a distributed ensemble model using precision-based weighted blending. Research on emotion recognition and classification refers to the detection of an individual’s emotional state by considering various types of data as input features, such as textual data, facial expressions, vocal, gesture and physiological signal recognition, electrocardiogram (ECG) and electrodermography (EDG)/galvanic skin response (GSR). The extraction of effective emotional features from different types of input data, as well as the analysis of large volume of real-time data, have become increasingly important tasks in order to perform accurate classification. Taking into consideration the volume and variety of the examined problem, a machine learning model that works in a distributed manner is essential. In this direction, we propose a precision-based weighted blending distributed ensemble model for emotion classification. The suggested ensemble model can work well in a distributed manner using the concepts of Spark’s resilient distributed datasets, which provide quick in-memory processing capabilities and also perform iterative computations effectively. Regarding model validation set, weights are assigned to different classifiers in the ensemble model, based on their precision value. Each weight determines the importance of the respective classifier in terms of its performing prediction, while a new model is built upon the derived weights. The produced model performs the task of final prediction on the test dataset. The results disclose that the proposed ensemble model is sufficiently accurate in differentiating between primary emotions (such as sadness, fear, and anger) and secondary emotions. The suggested ensemble model achieved accuracy of 76.2%, 99.4%, and 99.6% on the FER-2013, CK+, and FERG-DB datasets, respectively.

Evading text based emotion detection mechanism via adversarial attacks

Small and Medium-sized Enterprises (SMEs) in Sri Lanka are increasingly turning to social media to gain insights into consumer preferences; however, they frequently lack sophisticated analytical tools to derive actionable insights. This research introduces a hybrid deep learning framework aimed at sentiment and emotion analysis specifically for the Sri Lankan cake sector. Customer reviews and social media comments were gathered and processed through various methods including text cleaning, emoji translation, stopword elimination, tokenization, and sequence padding. Several deep learning models— Convolutional Neural Networks (CNN), Bidirectional Long Short-Term Memory (BiLSTM), and Gated Recurrent Units (GRU) were assessed, along with hybrid models that combined CNN and BiLSTM with GPT-2 embeddings. The CNN+GPT-2 model achieved the highest accuracy in sentiment classification at 96.14%, while BiLSTM performed best in emotion classification with an accuracy of 91.57%. In terms of trend forecasting, Long Short-Term Memory (LSTM) outshone both SARIMA and Prophet, yielding the lowest error metrics and the most reliable sentiment predictions. These results underscore the potential of hybrid deep learning and transformer-based models to effectively capture complex consumer sentiments and provide SMEs with practical, AI-driven insights. This study is among the first to implement such hybrid NLP models within a localized Sri Lankan context, presenting a scalable approach for business intelligence and informed decision-making.