Sentiment Analysis Task Research Articles

Lifelong Text-Audio Sentiment Analysis learning

Sentiment analysis refers to the mining of textual context, which is conducted with the aim of identifying and extracting subjective opinions in textual materials. However, most existing methods neglect other important modalities, e.g., the audio modality, which can provide intrinsic complementary knowledge for sentiment analysis. Furthermore, much work on sentiment analysis cannot continuously learn new sentiment analysis tasks or discover potential correlations among distinct modalities. To address these concerns, we propose a novel Lifelong Text-Audio Sentiment Analysis (LTASA) model to continuously learn text-audio sentiment analysis tasks, which effectively explores intrinsic semantic relationships from both intra-modality and inter-modality perspectives. More specifically, a modality-specific knowledge dictionary is developed for each modality to obtain shared intra-modality representations among various text-audio sentiment analysis tasks. Additionally, based on information dependence between text and audio knowledge dictionaries, a complementarity-aware subspace is developed to capture the latent nonlinear inter-modality complementary knowledge. To sequentially learn text-audio sentiment analysis tasks, a new online multi-task optimization pipeline is designed. Finally, we verify our model on three common datasets to show its superiority. Compared with some baseline representative methods, the capability of the LTASA model is significantly boosted in terms of five measurement indicators.

A novel text sentiment analysis system using improved depthwise separable convolution neural networks.

Human behavior is greatly affected by emotions. Human behavior can be predicted by classifying emotions. Therefore, mining people's emotional tendencies from text is of great significance for predicting the behavior of target groups and making decisions. The good use of emotion classification technology can produce huge social and economic benefits. However, due to the rapid development of the Internet, the text information generated on the Internet increases rapidly at an unimaginable speed, which makes the previous method of manually classifying texts one-by-one more and more unable to meet the actual needs. In the subject of sentiment analysis, one of the most pressing problems is how to make better use of computer technology to extract emotional tendencies from text data in a way that is both more efficient and accurate. In the realm of text-based sentiment analysis, the currently available deep learning algorithms have two primary issues to contend with. The first is the high level of complexity involved in training the model, and the second is that the model does not take into account all of the aspects of language and does not make use of word vector information. This research employs an upgraded convolutional neural network (CNN) model as a response to these challenges. The goal of this model is to improve the downsides caused by the problems described above. First, the text separable convolution algorithm is used to perform hierarchical convolution on text features to achieve the refined extraction of word vector information and context information. Doing so avoids semantic confusion and reduces the complexity of convolutional networks. Secondly, the text separable convolution algorithm is applied to text sentiment analysis, and an improved CNN is further proposed. Compared with other models, the proposed model shows better performance in text-based sentiment analysis tasks. This study provides great value for text-based sentiment analysis tasks.

Tourism Reputation Index for Assessing Perceptions on Destinations Using Collaborative Text Data

The tourism sector is experiencing a change in trend due to the widespread use of Web 2.0 by tourists. This phenomenon has prompted tourism agents to apply Big Data techniques and Business Intelligence tools to find out what tourists think. In this sense, the analysis of the Electronic Word Of Mouth in Social Networks is particularly relevant, and the academic literature has faced some difficulties in extracting data, determining its meaning and linking it to the destination. The aim of this work is to is to present the results that can be generated by the design of a tourism reputation index, through the analysis of data created collaboratively on the Twitter Social Network. For this purpose, a BI tool has been created to extract data from short text messages posted by users (tweets) and process them by implementing an Extract, Transform and Load process. In this process, Natural Language Processing is carried out, focusing on Sentiment Analysis tasks, applying an automatic classification model (Multinomial Naive Bayes) and then an automatic thematic categorisation of these texts depending on the words they contain. Subsequently, the data is analysed, calculating a global online reputation index based on sub-indexes of perception on different tourism aspects. These ratings are shown to users through an interactive dashboard, which allows comparison between different queries associated with a tourist location and a temporal period. This research helps tourism agents to better understand the public perception about destinations and make appropriate and well-informed decisions in real time, facilitating the intelligent management of data that enables the creation of smart tourism destinations with competitive, unique and sustainable advantages.

Few-shot Aspect Category Sentiment Analysis via Meta-learning

Existing aspect-based/category sentiment analysis methods have shown great success in detecting sentiment polarity toward a given aspect in a sentence with supervised learning, where the training and inference stages share the same pre-defined set of aspects. However, in practice, the aspect categories are changing rather than keeping fixed over time. Dealing with unseen aspect categories is under-explored in existing methods. In this article, we formulate a new few-shot aspect category sentiment analysis (FSACSA) task, which aims to effectively predict the sentiment polarity of previously unseen aspect categories. To this end, we propose a novel Aspect-Focused Meta-Learning (AFML) framework that constructs aspect-aware and aspect-contrastive representations from external knowledge to match the target aspect with aspects in the training set. Concretely, we first construct two auxiliary contrastive sentences for a given sentence with the incorporation of external knowledge, enabling the learning of sentence representations with a better generalization. Then, we devise an aspect-focused induction network to leverage the contextual sentiment toward a given aspect to refine the label vectors. Furthermore, we employ the episode-based meta-learning algorithm to train the whole network, so as to learn to generalize to novel aspects. Extensive experiments on multiple real-life datasets show that our proposed AFML framework achieves the state-of-the-art results for the FSACSA task.

Text classification in memristor-based spiking neural networks

Memristors, emerging non-volatile memory devices, have shown promising potential in neuromorphic hardware designs, especially in spiking neural network (SNN) hardware implementation. Memristor-based SNNs have been successfully applied in a wide range of applications, including image classification and pattern recognition. However, implementing memristor-based SNNs in text classification is still under exploration. One of the main reasons is that training memristor-based SNNs for text classification is costly due to the lack of efficient learning rules and memristor non-idealities. To address these issues and accelerate the research of exploring memristor-based SNNs in text classification applications, we develop a simulation framework with a virtual memristor array using an empirical memristor model. We use this framework to demonstrate a sentiment analysis task in the IMDB movie reviews dataset. We take two approaches to obtain trained SNNs with memristor models: (1) by converting a pre-trained artificial neural network (ANN) to a memristor-based SNN, or (2) by training a memristor-based SNN directly. These two approaches can be applied in two scenarios: offline classification and online training. We achieve the classification accuracy of 85.88% by converting a pre-trained ANN to a memristor-based SNN and 84.86% by training the memristor-based SNN directly, given that the baseline training accuracy of the equivalent ANN is 86.02%. We conclude that it is possible to achieve similar classification accuracy in simulation from ANNs to SNNs and from non-memristive synapses to data-driven memristive synapses. We also investigate how global parameters such as spike train length, the read noise, and the weight updating stop conditions affect the neural networks in both approaches. This investigation further indicates that the simulation using statistic memristor models in the two approaches presented by this paper can assist the exploration of memristor-based SNNs in natural language processing tasks.

Text Sentiment Classification Based on BERT Embedding and Sliced Multi-Head Self-Attention Bi-GRU.

In the task of text sentiment analysis, the main problem that we face is that the traditional word vectors represent lack of polysemy, the Recurrent Neural Network cannot be trained in parallel, and the classification accuracy is not high. We propose a sentiment classification model based on the proposed Sliced Bidirectional Gated Recurrent Unit (Sliced Bi-GRU), Multi-head Self-Attention mechanism, and Bidirectional Encoder Representations from Transformers embedding. First, the word vector representation obtained by the BERT pre-trained language model is used as the embedding layer of the neural network. Then the input sequence is sliced into subsequences of equal length. And the Bi-sequence Gated Recurrent Unit is applied to extract the subsequent feature information. The relationship between words is learned sequentially via the Multi-head Self-attention mechanism. Finally, the emotional tendency of the text is output by the Softmax function. Experiments show that the classification accuracy of this model on the Yelp 2015 dataset and the Amazon dataset is 74.37% and 62.57%, respectively. And the training speed of the model is better than most existing models, which verifies the effectiveness of the model.

Construction and Application of English-Chinese Multimodal Emotional Corpus Based on Artificial Intelligence

Multimodal corpus is a new type of multimedia teaching aid tool that was born and widely used in the current social development and educational reform process. It is mainly based on computer technology and network technology, and uses various multimedia materials as corpus to establish a more comprehensive English database. Corpus is increasingly used in English-Chinese multimodal teaching. Combining the content of the corpus with modern information technology will help to better understand the meaning, usage and collocation of English and Chinese multimodality, and then improve the initiative of autonomous learning. Furthermore, English-Chinese multimodality is a fixed or semi-fixed programmed language, which can be directly extracted from memory according to the context during use, which can effectively improve learners’ writing efficiency and overcome the negative impact of native language transfer. Facing the task of sentiment analysis, this article has taken the text semantic representation technology based on representation learning as the main method, and carries out a series of researches from the definition of implicit emotion, the characteristics of language expression and the recognition method. Finally, in the language system of modern Chinese, it is found that there are many sentence patterns that are very similar to the middle verb structure in language expression, but the efficiency is still improved by 3.6%. Compared with traditional methods, the English-Chinese multimodal sentiment corpus method based on artificial intelligence can better reflect the application value of language, which also helps to feel the real context, change the way of thinking, and make the way of thinking more suitable for the language awareness of native speakers, thereby improving the vividness of language expression.

GCAT-GTCU: Graph-Connected Attention Network and Gate Than Change Unit for Aspect-Level Sentiment Analysis

Currently, attention mechanisms are widely used in aspect-level sentiment analysis tasks. Previous studies have only used attention mechanisms combined with neural networks for aspect-level sentiment classification, and the feature extraction of the model is insufficient. When the same aspect and sentiment polarity appear in multiple sentences, the semantic information sharing of the same domain is also ignored, resulting in low model performance. To address these problems, the paper proposes an aspect-level sentiment analysis model, GCAT-GTCU, which combines a Graph-connected Attention Network containing symmetry with Gate Than Change Unit. Three nodes of words, sentences, and aspects are constructed, and local and deep-level features of sentences are extracted using CNN splicing BiGRU; node connection information is added to GAT to form a GCAT containing symmetry to realize the information interaction of three nodes, pay attention to the contextual information, and update the shared information of three nodes at any time; a new gating mechanism GTCU is constructed to filter noisy information and control the flow of sentiment information; finally, the three nodes are extracted information to predict the final sentiment polarity. The experimental results on four publicly available datasets show that the model outperforms the baseline model against which it is compared in some very controlled situations.

Sentiment enhanced answer generation and information fusing for product-related question answering

The reviews written by users in E-commerce platform have been fully exploited by product-related question answering systems, which ignore the product descriptions with valuable information to improve the system performance, especially for objective questions. In this paper, we propose a model named as Sentiment-enhanced Answer Generation and product Descriptions Fusing (SAGDF) for Product-related Question Answering task. The valid information from user reviews and product descriptions are learned jointly for answer generation. Importantly, the sentiment analysis task is performed initially to make the model judge the sentiment tendency correctly for answer accuracy. The review encoder of the model is fine-tuned on the large-scale Amazon review dataset to implement the sentiment analysis task. In addition, we build the dataset named AmazonQARD and it contains both reviews and corresponding product descriptions. The proposed model is trained on AmazonQARD and it processes user reviews individually and integrates product descriptions to constantly correct the generated answers iteratively. Experimental results demonstrate that SAGDF significantly outperforms the other baseline systems.

MBi-GRUMCONV: A novel Multi Bi-GRU and Multi CNN-Based deep learning model for social media sentiment analysis

Today, internet and social media is used by many people, both for communication and for expressing opinions about various topics in many domains of life. Various artificial intelligence technologies-based approaches on analysis of these opinions have emerged natural language processing in the name of different tasks. One of these tasks is Sentiment analysis, which is a popular method aiming the task of analyzing people’s opinions which provides a powerful tool in making decisions for people, companies, governments, and researchers. It is desired to investigate the effect of using multi-layered and different neural networks together on the performance of the model to be developed in the sentiment analysis task. In this study, a new, deep learning-based model was proposed for sentiment analysis on IMDB movie reviews dataset. This model performs sentiment classification on vectorized reviews using two methods of Word2Vec, namely, the Skip Gram and Continuous Bag of Words, in three different vector sizes (100, 200, 300), with the help of 6 Bidirectional Gated Recurrent Units and 2 Convolution layers (MBi-GRUMCONV). In the experiments conducted with the proposed model, the dataset was split into 80%-20% and 70%-30% training-test sets, and 10% of the training splits were used for validation purposes. Accuracy and F1 score criteria were used to evaluate the classification performance. The 95.34% accuracy of the proposed model has outperformed the studies in the literature. As a result of the experiments, it was found that Skip Gram has a better contribution to classification success.

Hoax COVID-19 News Detection Based on Sentiment Analysis in Indonesian using Support Vector Machine (SVM) Method

The increasing use of technology makes it easier for information media such as news to be disseminated and does not demand possibilities, there is a lot of hoax news spreading. Twitter is one of the media most frequently used by the public to access and disseminate information. This research will focus on detecting Indonesian language COVID-19 news taken from Twitter. Detection of hoax news can be assisted by using sentiment analysis, one of the uses of classification text. Support Vector Machine (SVM) can be used to perform sentiment analysis tasks. After getting the sentiment analysis results, the hoax detection process will use the Bag of Words. Bag of Words is a collection of word dictionaries for weighting words to determine specific labels. The built SVM model succeeded in classifying tweet repliessentiment with an average accuracy of 83.17% with a threshold of 35%. At the same time, the hoax detection process gets the best accuracy of 62.5% with a threshold of -5 or -6.

Deep Learning Using Context Vectors to Identify Implicit Aspects

Aspects extraction is the key task in the sentiment analysis problem, which includes extraction of both explicit and implicit aspects. Identifying implicit aspects is not a new task in sentiment analysis, but it still presents many challenges. Numerous studies have addressed the issue and offered different approaches, but there are still fundamental challenges as follows: solving the context word problem; the domain-specific problem; and the implicit aspects represented by sentiment words that do not appear explicitly in the text. This paper proposes a method for identifying implicit aspects of sentiment words. To address these above challenges, the system is built on the foundation of deep learning with context vectors, dependency grammar, anaphora coreference resolution, and sentiment ontology. The dependency parser and the entity coreference resolver are adopted to filter out the sentiment–aspect pairs data generated from embedding context words. Then, a fine-tuned technique is applied, and the classifier corresponding to each sentiment word is built to identify implicit aspects. This combination has produced positive results, reaching almost 90% accuracy.

An Experimental Analysis of Deep Neural Network Based Classifiers for Sentiment Analysis Task

The application of natural language processing (NLP) in sentiment analysis task by using textual data has wide scale application across various domains in plethora of industries. We have methodically studied pre-existing models and proposed new models for examining sentiment analysis task. The models proposed were analysed with three widely popular word embeddings separately and in combined approach using all embeddings as unique channels. We combined deep neural network models such as Bidirectional Long Short-Term Memory (BiLSTM) and Convolutional Neural Network (CNN) so that integrated models complement each other with their unique architectures. The word embeddings used had profound impact in accuracy of models owing to performative changes. The best word embedding was Word2Vec giving highest accuracy in almost all implemented models, followed by GloVe. FastText embedding performed consistently worse, giving much lower accuracy than other embeddings. We also observed that adding transformer encoder layers with CNN improves accuracy by 2% when compared to CNN without any transformer layers. An accuracy improvement of 2-3% over CNN-BiLSTM model was also observed by utilizing transformer encoder layer in conjunction with both BiLSTM and CNN. The proposed model achieved an accuracy of 89.04% on SST-2 dataset. We also compared larger pretrained language model used in sentiment analysis task with our proposed approach. The accuracy values obtained through combination of embeddings and models can be useful for other researchers when selecting word embeddings for their models.

Dense Concatenation Memory Network for Aspect Level Sentiment Analysis

Aspect-level sentiment analysis is a fundamental task in sentiment analysis. It has many practical applications in product reviews, takeaway evaluations, and public opinion supervision. Over the past few years, previous studies have made remarkable progress. However, the existing methods still have some areas for improvement in aspect processing and feature fusion. In this study, we propose an aspect-context dense concatenation model (ACDC) to merge deep semantic information from various aspects and contexts. ACDC first captures the interaction information between the aspect and context using an interaction matrix. Subsequently, a dense concatenation mechanism is used to extract and integrate features at different levels. In addition, we introduce two loss functions to eliminate the interference information and overfitting. On the Restaurant2014 and Laptop2014 datasets, the experimental results show that our proposed model achieves state-of-the-art performance in terms of accuracy and MF1. Furthermore, comparative experiments are conducted to verify the effectiveness of the proposed model.

A Comprehensive Survey on Arabic Sarcasm Detection: Approaches, Challenges and Future Trends

On social media platforms, it is essential to express one’s thoughts, opinions, and reviews. One of the most widely used linguistic forms to criticize or express a person’s ideas with ridicule is sarcasm, where the written text has both intended and unintended meanings. The sarcastic text frequently reverses the polarity of the sentiment. Therefore, detecting sarcasm in the text has a positive impact on the sentiment analysis task and ensures more accurate results. Although Arabic is one of the most frequently used languages for web content sharing, the sarcasm detection of Arabic content is restricted and yet still naive due to several challenges, including the morphological structure of the Arabic language, the variety of dialects, and the lack of adequate data sources. Despite that, researchers started investigating this area by introducing the first Arabic dataset and experiment for irony detection in 2017. Thus, our review focuses on studies published between 2017 and 2022 on Arabic sarcasm detection. We provide a thorough literature review of Artificial Intelligence (AI) techniques and benchmarks used for Arabic sarcasm detection. In addition, the challenges of Arabic sarcasm detection are investigated, along with future directions, focusing on the challenge of publicly available Arabic sarcasm datasets.

Impact of Vector Embeddings on the Performance of Tolerance Near Sets-based Sentiment Classifier for Text Classification

In recent years, Natural Language Processing (NLP) has gained significant attention, and sentiment analysis is an essential subfield of NLP that deals with identifying the sentiment or emotion conveyed in the text. Tolerance near sets (TNS) is a mathematical framework that has shown promising results in sentiment analysis tasks. However, the choice of word embeddings can significantly impact the performance of TNS-based classifiers. This paper investigates the impact of using different embeddings on the performance of tolerance near sets-based sentiment classifiers. This paper compares the use of different embeddings, including DistilBERT, MiniLM, and Word Embeddings, and their combinations, to understand their impact on TNS-based sentiment analysis. The TSC 2.0 model proposed in this paper achieves a weighted F1 score of 92.1% in one of the datasets, an improvement due to the sentence embeddings used. Experimental results have led to the observation that tie-breaking and variance-based classification may have led to a noticeable improvement in cases with more than three.

Multimodal Information Bottleneck: Learning Minimal Sufficient Unimodal and Multimodal Representations

Learning effective joint embedding for cross-modal data has always been a focus in the field of multimodal machine learning. We argue that during multimodal fusion, the generated multimodal embedding may be redundant, and the discriminative unimodal information may be ignored, which often interferes with accurate prediction and leads to a higher risk of overfitting. Moreover, unimodal representations also contain noisy information that negatively influences the learning of cross-modal dynamics. To this end, we introduce the multimodal information bottleneck (MIB), aiming to learn a powerful and sufficient multimodal representation that is free of redundancy and to filter out noisy information in unimodal representations. Specifically, inheriting from the general information bottleneck (IB), MIB aims to learn the minimal sufficient representation for a given task by maximizing the mutual information between the representation and the target and simultaneously constraining the mutual information between the representation and the input data. Different from general IB, our MIB regularizes both the multimodal and unimodal representations, which is a comprehensive and flexible framework that is compatible with any fusion methods. We develop three MIB variants, namely, early-fusion MIB, late-fusion MIB, and complete MIB, to focus on different perspectives of information constraints. Experimental results suggest that the proposed method reaches state-of-the-art performance on the tasks of multimodal sentiment analysis and multimodal emotion recognition across three widely used datasets. The codes are available at \url{https://github.com/TmacMai/Multimodal-Information-Bottleneck}.

HIM: An End-to-End Hierarchical Interaction Model for Aspect Sentiment Triplet Extraction

Aspect Sentiment Triplet Extraction (ASTE) is an emerging task of fine-grained sentiment analysis, which aims to extract aspect terms, associated opinion terms, and sentiment polarities in the form of triplets. Thus, ASTE involves two groups of subtasks: aspect/opinion term extraction and aspect-opinion-pair sentiment classification. Due to the high correlations of subtasks, three categories of joint methods have been proposed, including <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">end-to-end tagging-based methods</i> , <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">cascaded span-based methods</i> , and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">sequence-to-sequence generation-based methods</i> . These methods basically learn either a shared feature space or a shared sentence encoder to capture interactions across all subtasks by parameter sharing. However, they fail to learn deep and mutual interactive features for ASTE. In this work, we present a novel tagging scheme to cast ASTE as a unified boundary-words relation classification problem. Subsequently, we propose an end-to-end Hierarchical Interaction Model (HIM), exploiting deep and mutual interactions across subtasks mainly with two interaction modules. The first-level interaction module primarily leverages multi-task learning models to capture implicit subtask interactions. Then, the second-level interaction module, namely Gated Interaction Network (GIN), adopts a novel gated control mechanism and a newly-designed Conditional BiLSTM (Cond-BiLSTM) network to capture explicit subtask interactions. Moreover, to refine the unreliable outputs of the first-level module, we develop a General word-Pair Relationship Learning (G-PRL) component. With the task-shared features as input, G-PRL further facilitates interactions between term extraction and pair classification. We conduct experiments on two benchmarks and achieve promising results. Extensive analyses demonstrate the effectiveness and flexibility of our work.

Battle Royale Optimization with Fuzzy Deep Learning for Arabic Sentiment Classification

Aspect-Based Sentiment Analysis (ABSA) on Arabic corpus has become an active research topic in recent days. ABSA refers to a fine-grained Sentiment Analysis (SA) task that focuses on the extraction of the conferred aspects and the identification of respective sentiment polarity from the provided text. Most of the prevailing Arabic ABSA techniques heavily depend upon dreary feature-engineering and pre-processing tasks and utilize external sources such as lexicons. In literature, concerning the Arabic language text analysis, the authors made use of regular Machine Learning (ML) techniques that rely on a group of rare sources and tools. These sources were used for processing and analyzing the Arabic language content like lexicons. However, an important challenge in this domain is the unavailability of sufficient and reliable resources. In this background, the current study introduces a new Battle Royale Optimization with Fuzzy Deep Learning for Arabic Aspect Based Sentiment Classification (BROFDL-AASC) technique. The aim of the presented BROFDL-AASC model is to detect and classify the sentiments in the Arabic language. In the presented BROFDL-AASC model, data pre-processing is performed at first to convert the input data into a useful format. Besides, the BROFDL-AASC model includes Discriminative Fuzzy-based Restricted Boltzmann Machine (DFRBM) model for the identification and categorization of sentiments. Furthermore, the BRO algorithm is exploited for optimal fine-tuning of the hyperparameters related to the FBRBM model. This scenario establishes the novelty of current study. The performance of the proposed BROFDL-AASC model was validated and the outcomes demonstrate the supremacy of BROFDL-AASC model over other existing models.

An Extreme Gradient Boosting Approach for Classification and Sentiment Analysis

Since 2020, when the coronavirus epidemic was at its peak, the Indonesian Ministry of Health's social media accounts have been constantly followed by a big number of individuals. The Indonesian Ministry of Health account is a fantastic resource for social media users, particularly Twitter users. The Republic of Indonesia's Ministry of Health's Twitter account publishes a wide range of content at random. As a result, it is usually difficult for Twitter users to determine the type of information provided by the Ministry of Health of the Republic of Indonesia. The positive and negative responses of Twitter users to material released by the Indonesian Ministry of Health's Twitter account are frequently noted. The decision tree algorithm is tree-based, similar to the extreme gradient boosting method (XGBoost). The extreme gradient boosting approach has been successfully implemented with high performance in the classification process. This classification is separated into two primary categories: general and essential information categorization and sentiment analysis, which is classified into three classes: positive, neutral, and negative. Both the classification work and the sentiment analysis produced outstanding accuracy levels. Based on 2243 tweets, an accuracy rate of 89.35% has been achieved for classification, supported by a precision of 88.76% and a recall value of 88.58% when using 80 data training and 20 data testing. Similarly, the maximum accuracy in sentiment analysis was achieved utilizing the same 80-20 data partitioning, with a 91.22% accuracy rate. Using 304 comments data, accuracy was calculated to be 89.17% and recall was calculated to be 89.06%. It's worth noting that an 80-20 split for training and testing consistently produced the best results for both the sentiment analysis and classification tasks.