A comprehensive dataset for sentiment and emotion classification from Bangladesh e-commerce reviews

In the rapidly evolving domain of e-commerce, analyzing customer feedback through reviews is crucial, particularly for understanding and enhancing consumer experience in the Bangladeshi market. Our comprehensive dataset, derived from two Bangladeshi e-commerce platforms, Daraz and Pickaboo, features a diverse collection of reviews in both Bengali and English, covering a broad range of products. These reviews are not only rich in linguistic variety but also encapsulate a spectrum of emotions, some even conveyed through emojis, offering a deep dive into consumer sentiment. Expert annotators have meticulously examined and categorized each review, classifying emotions into five distinct types - Happiness, Sadness, Fear, Anger, and Love - and sentiments into Positive (Happiness, Love) and Negative (Sadness, Anger, Fear) categories. This level of detailed annotation enables precise assessments of customer emotions and preferences, which are essential for evaluating and improving existing product offerings. Moreover, the insights gleaned from this dataset are invaluable for guiding future product development and uncovering new opportunities in the dynamic Bangladeshi market. Ultimately, this dataset not only serves as a significant resource for sentiment analysis using natural language processing (NLP) techniques but also contributes valuable insights into the unique consumer behavior patterns in Bangladesh, enriching the NLP community's understanding of diverse market dynamics.

Multilabel multiclass sentiment and emotion dataset from indonesian mobile application review

Reviews are a person's way of expressing feedback on something in the form of criticism and ideas. Reviews of mobile apps are a type of user feedback that focuses on the performance and look of a mobile application and is typically featured on the download page of a mobile application, such as in the Apps Store. Because it comprises a person's feelings and emotions, whether they are joyful, sad, hostile, or indifferent toward a mobile application, the review data is textual and may be gathered and utilized as material for creating a textual dataset. This work creates a multi-label multi-class Indonesian-language dataset based on public reviews of mobile applications with sentiment and emotional values. Another factor supporting the creation of this dataset is the fact that there is still a limited number of textual datasets based on the Indonesian language that are multi-label multiclass for performing sentiment analysis tasks, particularly those linked to text classification tasks. The data generated by this research was cleaned and handled during the pre-processing step and was annotated with 3 sentiments, namely positive, negative, and neutral, as well as 6 emotions, namely anger, fear, sad, happy, love, and neutral.

Multilabel convolution neural network for facial expression recognition and ordinal intensity estimation

Facial Expression Recognition (FER) has gained considerable attention in affective computing due to its vast area of applications. Diverse approaches and methods have been considered for a robust FER in the field, but only a few works considered the intensity of emotion embedded in the expression. Even the available studies on expression intensity estimation successfully assigned a nominal/regression value or classified emotion in a range of intervals. Most of the available works on facial expression intensity estimation successfully present only the emotion intensity estimation. At the same time, others proposed methods that predict emotion and its intensity in different channels. These multiclass approaches and extensions do not conform to man heuristic manner of recognising emotion and its intensity estimation. This work presents a Multilabel Convolution Neural Network (ML-CNN)-based model, which could simultaneously recognise emotion and provide ordinal metrics as the intensity estimation of the emotion. The proposed ML-CNN is enhanced with the aggregation of Binary Cross-Entropy (BCE) loss and Island Loss (IL) functions to minimise intraclass and interclass variations. Also, ML-CNN model is pre-trained with Visual Geometric Group (VGG-16) to control overfitting. In the experiments conducted on Binghampton University 3D Facial Expression (BU-3DFE) and Cohn Kanade extension (CK+) datasets, we evaluate ML-CNN's performance based on accuracy and loss. We also carried out a comparative study of our model with some popularly used multilabel algorithms using standard multilabel metrics. ML-CNN model simultaneously predicts emotion and intensity estimation using ordinal metrics. The model also shows appreciable and superior performance over four standard multilabel algorithms: Chain Classifier (CC), distinct Random K label set (RAKEL), Multilabel K Nearest Neighbour (MLKNN) and Multilabel ARAM (MLARAM).