Sentiment Analysis

Today, social networks are fast and dynamic communication intermediaries that are a vital business tool. This study aims at examining the views of those involved with Facebook stocks so that we can summarize their views to predict the general behavior of this stock and collectively consider possible Facebook stock price movements, and create a more accurate pattern compared to previous patterns. In this study, we have analyzed two statistical samples, the first being a large dataset containing a variety of tweets with an emotional tag. That is, it needed a set that had already been extracted from each individual tweet by a trusted human or machine. Consequently, we have collected posts on Facebook in an eighty-day period. In this study, we used a tagged dataset using Python's programming language and vector-to-word algorithm. The research results show that, we need stock change information, machine learning and sentiment analysis, and on paper we conclude that positive news about a company excites people to have positive opinions about it which in turn results in people encouraging each other to buy and hold stocks. Meanwhile, the opposite trend is also true, but everything will not always be easy and clear, and it is in areas of high complexity and mental uncertainty that the art of using the three elements mentioned above is evident.

Social media allows people interact to express their thoughts or feelings about different subjects. However, some of users may write offensive twits to other via social media which known as cyber bullying. Successful prevention depends on automatically detecting malicious messages. Automatic detection of bullying in the text of social media by analyzing the text "twits" via one of the machine learning algorithms. In this paper, we have reviewed algorithms for automatic cyberbullying detection in Arabic of machine learning, and after comparing the highest accuracy of these classifications we will propose the techniques Ridge Regression (RR) and Logistic Regression (LR), which achieved the highest accuracy between the various techniques applied in the automatic cyberbullying detection in English and between the techniques that was used in the sentiment analysis in Arabic text, The purpose of this work is applying these techniques for detecting cyberbullying in Arabic.

Online reviews have become essential aspect in E-commerce platforms due to its role for assisting customers’ buying choices. Furthermore, the most helpful reviews that have some attributes are support customers buying decision; therefore, there is needs for investigating what are the attributes that increase the Review Helpfulness (RH). This research paper proposed novel model called inclusive review helpfulnessmodel (IRHM) can be used to detect the most attributes affecting the RH and build classifier that can predict RH based on these attributes. IRHM is implemented on Amazon.com using collection of reviews from different categories. The results show that IRHM can detect the most important attributes and classify the reviews as helpful or not with accuracy of 94%, precision of 0.20 and had excellent area under curve close to 0.94.

Sentiment analysis is the process of analyzing a person’s perception or belief about a particular subject matter. However, finding correct opinion or interest from multi-facet sentiment data is a tedious task. In this paper, a method to improve the sentiment accuracy by utilizing the concept of categorized dictionary for sentiment classification and analysis is proposed.  A categorized dictionary is developed for the sentiment classification and further calculation of sentiment accuracy. The concept of categorized dictionary involves the creation of dictionaries for different categories making the comparisons specific. The categorized dictionary includes words defining the positive and negative sentiments related to the particular category. It is used by the mapper reducer algorithm for the classification of sentiments. The data is collected from social networking site and is pre-processed. Since the amount of data is enormous therefore a reliable open-source framework Hadoop is used for the implementation. Hadoop hosts various software utilities to inspect and process any type of big data. The comparative analysis presented in this paper proves the worthiness of the proposed method.

Brand community interactions and online customer support have become major platforms of brand sentiment strengthening and loyalty creation. Rapid brand responses to each customer request though inbound tweets in twitter and taking proper actions to cover the needs of customers are the key elements of positive brand sentiment creation and product or service initiative management in the realm of intense competition. In this research, there has been an attempt to collect near three million tweets of inbound customer requests and outbound brand responses of international enterprises for the purpose of brand sentiment analysis. The steps of CRISP-DM have been chosen as the reference guide for business and data understanding, data preparation, text mining, validation of results as well as the final discussion and contribution. A rich phase of text pre-processing has been conducted and various algorithms of sentiment analysis were applied for the purpose of achieving the most significant analytical conclusions over the sentiment trends. The findings have shown that the sentiment of customers toward a brand is significantly correlated with the proper response of brands to the brand community over social media as well as providing the customers with a deep feeling of reciprocal understanding of their needs in a mid-to-long range planning.

Recommender systems are important tools for users to identify their preferred items and for businesses to improve their products and services. In recent years, the use of online services for selection and reservation of hotels have witnessed a booming growth. Customer’ reviews have replaced the word of mouth marketing, but searching hotels based on user priorities is more time-consuming. This study is aimed at designing a recommender system based on the explicit and implicit preferences of the customers in order to increase prediction’s accuracy. In this study, we have combined sentiment analysis with the Collaborative Filtering (CF) based on deep learning for user groups in order to increase system accuracy. The proposed system uses Natural Language Processing (NLP) and supervised classification approach to analyze sentiments and extract implicit features. In order to design the recommender system, the Singular Value Decomposition (SVD) was used to improve scalability. The results show that our proposed method improves CF performance.

In attempts to examine the mapped spaces of a literary narrative, various quantitative approaches have been deployed to extract data from texts to graphs, maps, and trees. Though the existing methods offer invaluable insights, they undertake a rather different project than that of literary scholars who seek to examine privileged or unprivileged representations of certain spaces. This study aims to propose a computerized method to examine how matters of space and spatiality are addressed in literary writings. As the primary source of data, the study will focus on Viet Thanh Nguyen’s The Sympathizer (2015), which explores the lives of Vietnamese diaspora in two geographical locations, Vietnam, and America. To examine the portrayed spatial relations, that is which country is privileged over the other, and to find out the underlying opinion about the two places, this study performs topic modelling with Latent Dirichlet Allocation (LDA) and Latent Semantic Analysis (LSA) by using TextBlob. In addition, Python is used as the analytical tool for this project as it supports two LDA algorithms: Gensim and Mallet. To overcome the limitation that the performance of the model relies on the available libraries in Python, the study employs machine learning approach. Even though the results indicated that both geographical spaces are portrayed slightly positively, America achieves a higher polarity score than Vietnam and hence seems to be the favored space in the novel. This study can assist literary scholars in analyzing spatial relations more accurately in large volumes of works.

The diversity and high volume of available information on the web make data retrieval a serious challenge in this environment. On the other hand, obtaining user satisfaction is difficult, which is one of the main challenges of data retrieval systems. Depending on their information about interests and needs for the same keyword, different people expect different responses from Information Retrieval (IR) systems. Achieving this goal requires an effective method to retrieve information. Personalized Information Retrieval (PIR) is an effective method to achieve this goal which is considered by researchers today.  Folksonomy is the process that allows users to tag in a specific domain of information in a social environment (tags are accessible to other users). Folksonomy systems are made collaborative tagging systems. Due to the large volume and variety of tags produced, resolving ambiguity is a severe challenge in these systems. In recent years, word embedding methods have been considered by researchers as a successful method to fix the ambiguity of texts. This study proposes a model which, in addition to using word embedding methods to remove tag ambiguity, provides search results in a personalized approach by fixing ambiguity and sentiment analysis combination tailored to users' interests. In this research, different models of word embeddings were applied. The experiments' results show that after applying the fixing ambiguity, the mean accuracy criterion improved by 1.93% and the mean MRR ( Mean Reciprocal Rank)  by 0.38%.

With the fast growth of social media, they have become the most important platform for posting multimodal content generated by users. Much of the data on social networks such as Instagram and Telegram is multimodal data. With the aim of analyzing such multimodal data in social networks, multimodal sentiment analysis has become one of the most significant subjects for researchers in the field of emotion recognition and data mining. Although multimodal sentiment analysis of social media data for English language has been addressed in several researches recently, few studies addressed the problem for the Persian language which is the official language of more than 120 million of people around the word. In this study, a multimodal deep learning model is proposed to address this problem. The proposed method utilizes a bi-directional long short-term memory (bi-LSTM) for processing text posts and a VGG16 convolutional network for analyzing images. A new dataset of Instagram and Telegram posts, MPerSocial, containing 1000 pairs of images and Persian comments is introduced in the current study and used for evaluating the proposed method. The results of experiments show that using the fusion of textual and image modalities improves sentiment polarity detection accuracy by 20% and 8% compared with the scenario in which image and text modalities in isolation. Also, the performance of the proposed model is better than three similar deep and four traditional machine learning models. All codes and dataset used in the current study are publicly available at GitHub.  

Sentiment analysis, as one of the text-mining techniques, has been widely used in many research areas related to public opinions. The present work seeks to examine the reception of the translations of  Louis Cha’s martial arts novels in the English and French worlds through sentiment analyses of readers’ online book reviews in popular book reviewing websites such as Amazon and Belio. The results show that the English and French versions of Louis Cha’s martial arts novels are mostly well received by readers since both English and French readers had given the highest ratings to the narrative features, plots and characterization features of the translated novels. Furthermore, between the translations in English and French, the reception level of the Anna Holmwood’s English version is higher than that of the French version by Jiann-yuh Wang. Overall, the findings of the study suggest that translations maintaining original Chinese cultural elements have been well-received in the Western community.

Social media is a popular data source in the research community. It provides different opportunities to design practical applications to favor humanity and society. A significant amount of people consumes social media content. Thus, sometimes content promoters and influencers publish misleading and toxic content. Therefore, this paper proposes an unhealthy content filtering system using the information retrieval model SOIR to identify and remove poisonous content from social media. The Semantic query Optimization-based Information Retrieval (SOIR) uses Fuzzy C Means (FCM) clustering to produce a particular data structure. To incorporate a query generation technique for the generation of multiple queries to increase the probability of correct outcomes. The SOIR model is modified in this work to utilize the model with the social media toxic content filtering model. The model uses linguistic and semantically information to craft new feature sets. The Part of Speech (POS) tagging is used to construct the linguistic feature. Finally, the pattern-matching algorithm is designed to classify the tweets as toxic or nontoxic. Based on lexical and semantic analysis of similar semantic queries (Tweets), it is identified with the class labels of the tweets. Twitter text posts are used to create training and test samples in this context. Here, a total of 2002 tweets are used for the experiment. The experimental study has been carried out with the different I.R. models (K-NN, Cosine) based on precision, recall, and F1-Score demonstrating the superiority of the proposed classification model

What is clear is that judges usually judge cases based on their knowledge, experience, personality, and sentiment. Due to high pressures and stress, it may be difficult for them to carefully examine documents and evidence, which leads to more subjective judgments. Legal judgment prediction with artificial intelligence algorithms can benefit judicial bodies, legal experts, and litigants as well as judges. In this research, we are looking at predicting legal sentences in drug cases involving the purchase, possession, concealment, or transportation of illicit drugs, using machine learning methods, and the effect of sentiment and emotions in case texts on predicting the severity of whipping, fines, and imprisonment. So, the text documents of 6000 Persian drug-related cases were pre-processed and then the translation of the NRC Glossary of Emotions and sentiment was used to give each item a score for positive or negative sentiment and a score for emotion. Then machine learning methods were used for modeling. BERT, TFIDF+Adaboost, and Skipgram+LSTM+CNN methods had the highest accuracy, respectively. Also, evaluation criteria were analyzed in situations where sentiment scores, emotional scores, or both were used in the prediction process along with judicial texts. Finally, it was found that the use of sentiment and emotion scores improves the accuracy of legal judgment predictions for all three types of sentences and that sentiments have a greater impact on the accuracy of legal judgment predictions than emotions

Internet-based social channels have turned into an important information repository for many people to get an idea about current trends and events happening around the world. As a result of Abundance of raw information on these social media platforms, it has become a crucial platform for businesses and individuals to make decisions based on social media analytics. The ever-expanding volume of online data available on the global network necessitates the use of specialized techniques and methods to effectively analyse and utilize this vast amount of information. This study's objective is to comprehend the textual information at the Lexical and Semantic level and to extract sentiments from this information in the most accurate way possible. To achieve this, the paper proposes to cluster semantically related words by evaluating their lexical similarity with respect to feature and sequence vectors. The proposed method utilizes Natural Language Processing, semantic and lexical clustering and hybrid C4.5 algorithm to extract six subcategories of emotions over three classes of sentiments based on word-based analysis of text. The proposed approach has yielded superior results with seven existing approaches in terms of parametric values, with an accuracy of 0.96, precision of 0.92, sensitivity of 0.94, and an f1-score of 0.92.

Sentiment analysis is the process of identifying and categorizing people’s emotions or opinions regarding various topics. The analysis of Twitter sentiment has become an increasingly popular topic in recent years. In this paper, we present several machine learning and a deep learning model to analysis sentiment of Persian political tweets. Our analysis was conducted using Bag of Words and ParsBERT for word representation. We applied Gaussian Naive Bayes, Gradient Boosting, Logistic Regression, Decision Trees, Random Forests, as well as a combination of CNN and LSTM to classify the polarities of tweets. The results of this study indicate that deep learning with ParsBERT embedding performs better than machine learning. The CNN-LSTM model had the highest classification accuracy with 89 percent on the first dataset and 71 percent on the second dataset. Due to the complexity of Persian, it was a difficult task to achieve this level of efficiency. The main objective of our research was to reduce the training time while maintaining the model's performance. As a result, several adjustments were made to the model architecture and parameters. In addition to achieving the objective, the performance was slightly improved as well.

With the advent of user-generated text information on the Internet, text sentiment analysis plays an essential role in online business transactions. The expression of feelings and opinions depends on the domains, which have different distributions. In addition, each of these domains or so-called product groups has its vocabulary and peculiarities that make analysis difficult. Therefore, different methods and approaches have been developed in this area. However, most of the analysis involved a single-domain and few studies on cross-domain mood classification using deep neural networks have been performed. The aim of this study was therefore to examine the accuracy and transferability of deep learning frameworks for the cross-domain sentiment analysis of customer ratings for different product groups as well as the cross-domain sentiment classification in five categories “very positive”, “positive”, “neutral”, “negative” and “very negative”. Labels were extracted and weighted using the Long Short-Term Memory (LSTM) Recurrent Neural Network. In this study, the RNN LSTM network was used to implement a deep transfer learning framework because of its significant results in sentiment analysis. In addition, two different methods of text representation, BOW and CBOW were used. Based on the results, using deep learning models and transferring weights from the source domain to the target domain can be effective in cross-domain sentiment analysis.

Contemporary machine learning models, like deep neural networks, require substantial labeled datasets for proper training. However, in areas such as natural language processing, a shortage of labeled data can lead to overfitting. To address this challenge, data augmentation, which involves transforming data points to maintain class labels and provide additional valuable information, has become an effective strategy. In this paper, a deep reinforcement learning-based text augmentation method for sentiment analysis was introduced, combining reinforcement learning with deep learning. The technique uses Deep Q-Network (DQN) as the reinforcement learning method to search for an efficient augmentation strategy, employing four text augmentation transformations: random deletion, synonym replacement, random swapping, and random insertion. Additionally, various deep learning networks, including CNN, Bi-LSTM, Transformer, BERT, and XLNet, were evaluated for the training phase. Experimental findings show that the proposed technique can achieve an accuracy of 65.1% with only 20% of the dataset and 69.3% with 40% of the dataset. Furthermore, with just 10% of the dataset, the method yields an F1-score of 62.1%, rising to 69.1% with 40% of the dataset, outperforming previous approaches. Evaluation on the SemEval dataset demonstrates that reinforcement learning can efficiently augment text datasets for improved sentiment analysis results.

The proliferation of e-commerce has led to an overwhelming volume of customer reviews, posing challenges for consumers who seek reliable product evaluations and for businesses concerned with the integrity of their online reputation. This study addresses the critical problem of detecting fake reviews by developing a comprehensive framework that integrates Natural Language Processing (NLP) and machine learning techniques. Our methodology centers on sentiment analysis to discern the emotional valence of reviews, coupled with Part-of-Speech (PoS) tagging to analyze linguistic patterns that may signal deception. We meticulously extract a rich set of textual and statistical features, providing a robust basis for our predictive models. To enhance classification performance, we strategically employ both traditional machine learning algorithms and powerful ensemble techniques. Experimental results underscore the efficacy of our approach in detecting fraudulent reviews. We achieved a notable F1-Score of 82.9% and an accuracy of 82.6%, demonstrating the potential to safeguard consumers from misleading information and protect businesses from unfair practices.