Deep learning based sentiment analysis and offensive language identification on multilingual code-mixed data Scientific Reports

This figure provides a clearer illustration of the nuanced differences between the Lin Similarity distributions of CT and CO than a boxplot. The value range of Wu-Palmer Similarity is divided into 10 subintervals, and the number of texts in CT and CO that fall into each subinterval is counted. This figure provides a clearer illustration of the nuanced differences between the Wu-Palmer Similarity distributions of CT and CO than a boxplot. This study was financially supported by the Major S&T project (Innovation 2030) of China(2021ZD ), Xi’an Major Scientific and Technological Achievements Transformation and Industrialization Project(20KYPT ). The left neighbor entropy, right neighbor entropy are calculated as shown in (2) and (3).

This proactive approach can improve customer satisfaction, loyalty and brand reputation. Finding the right tone on social media can be challenging, but sentiment analysis can guide you. Brands like MoonPie have found success by engaging in humorous and snarky interactions, increasing their positive mentions and building buzz. By analyzing how users interact with your content, you can refine your brand messaging to better resonate with your audience. Sprout Social is an all-in-one social media management platform that gives you in-depth social media sentiment analysis insights.

In recent years, NLP has become a core part of modern AI, machine learning, and other business applications. Even existing legacy apps are integrating NLP capabilities into their workflows. Incorporating the best NLP software into your workflows will help you maximize several NLP capabilities, including automation, data extraction, and sentiment analysis.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author. Sentiment analysis helps you gain insights into customer feedback, brand perception, or public opinion to improve on your business’s weaknesses and expand on its strengths. Then, benchmark sentiment performance against competitors and identify emerging threats.

• Moreover, the system can prioritize or flag urgent requests and route them to the respective customer service teams for immediate action with semantic analysis.
• Looks like the average sentiment is the most positive in world and least positive in technology!
• These studies collectively underline the evolution of Amharic sentiment analysis and its challenges, providing valuable insights for future research.
• The question of whether translational language should be regarded as a distinctive language variant has since sparked considerable debate in the field of translation studies.
• Therefore, all points above the decision boundary (diagonal blue line) have positive S3 and are then predicted to have a positive sentiment, and all points below the boundary have negative S3 and are thus predicted to have a negative sentiment.

Understanding how people feel about your business is crucial, but knowing their sentiment toward your competitors can provide a competitive edge. Social media sentiment analysis can help you understand why customers might prefer a competitor’s product over yours, allowing you to identify gaps and opportunities in your offerings. Sentiment analysis helps brands keep a closer eye on the emotions behind their social messages and mentions, ensuring they are more attentive to comments and concerns as they pop up. Addressing these conversations—both negative and positive—signals that you’re actively listening to your customers. The insights you gain from sentiment analysis can translate directly into positive changes for your business. By understanding and acting on these insights, you can enhance customer satisfaction, boost engagement and improve your overall brand reputation.

This involves identifying sentiment-indicative terms within these mentions and categorizing them as positive, negative‌ or neutral. Rather than focusing on a one-off compliment or complaint, brands should look at the bigger picture of their audience’s feelings. For example, a flurry of praise is definitely a plus and should be picked up in social sentiment analytics.

Inshorts, news in 60 words !

The preceding function shows us how we can easily convert accented characters to normal English characters, which helps standardize the words in our corpus. Often, unstructured text contains a lot of noise, especially if you use techniques like web or screen scraping. HTML tags are typically one of these components which don’t add much value towards understanding and analyzing text.

We also looked at the cross-correlation of the target series with our predictors (i.e., ERKs series) to see if they were in phase (positive signs of cross-correlation) or out of phase (negative sign)60,61. Sentiment analysis tools are essential to detect and understand customer feelings. Companies that use these tools to understand how customers feel can use it to improve CX. Companies can use customer sentiment to alert service representatives when the customer is upset and enable them to reprioritize the issue and respond with empathy, as described in the customer service use case. Sentiment analysis software notifies customer service agents — and software — when it detects words on an organization’s list.

• As with any supervised learning task, the data is first divided into features (Feed) and label (Sentiment).
• In addition to empirical research, scholars have recognized the importance of exploring alternative sources to gain a more comprehensive understanding of sexual harassment in the region.
• This hybrid model outperforms previous models, and when looking at the marginal differences between training, validation, and testing, the difference is small, showing how well the model works in unknown datasets and its generalization ability.

One-hot encoding of a document corpus is a vast sparse matrix resulting in a high dimensionality problem28. One more great choice for sentiment analysis is Polyglot, which is an open-source Python library used to perform a wide range of NLP operations. The library is based on Numpy and is incredibly fast while offering a large variety of dedicated commands. Idiomatic is an ideal choice for users who need to improve their customer experience, as it goes beyond the positive and negative scores for customer feedback and digs deeper into the root cause. It also helps businesses prioritize issues that can have the greatest impact on customer satisfaction, allowing them to use their resources efficiently. SAP HANA Sentiment Analysis is ideal for analyzing business data and handling large volumes of customer feedback, support tickets, and internal communications with other SAP systems.

Language Transformers

For our daily analysis, we aggregate sentiment scores captured from all tweets on day t to access its impact on the stock market performance in the coming t+1 day. For instance, we aggregate sentiment captured from tweets on July 10 to analyze the correlation between sentiment on the 10th/11th July and market volatility and returns. By highlighting these contributions, this study demonstrates the novel aspects of this research and its potential impact on sentiment analysis and language translation. Machine learning models such as reinforcement learning, transfer learning, and language transformers drive the increasing implementation of NLP systems.

The model had been trained using 20 epochs and the history of the accuracy and loss had been plotted and shown in Fig. To avoid overfitting, the 3 epochs were chosen as the final model, where the prediction accuracy is 84.5%. However, its low recall for physical sexual harassment results in an F1 score of 60%, which represents the harmonic mean of precision and recall.

CNN-1D is mostly utilized in computer vision, but it also excels at classification problems in the natural language processing field. A CNN-1D is particularly capable If you intend to obtain new attributes from brief fixed-length chunks of the entire data set and the position of the feature is irrelevant62,63. Deep learning-based approach for danmaku sentiment analysis by multilayer neural networks. Li et al.35 used the XLNet model to evaluate the overall sentiment of danmaku comments as pessimistic or optimistic. Kapočiūtė-Dzikienė et al.29, claim that deep learning models tend to underperform when used for morphologically rich languages and hence recommend traditional machine learning approach with manual feature engineering. Despite the author’s conclusion, the recommendation does not hold true when comparing the performance of Amharic sentiment analysis model constructed in this study using deep learning with machine learning model proposed by Refs.6, 18.

The inclusion of external experts to validate the selection of keywords is aligned with the methodology used in similar studies39. These keywords provide insight into the concerns and priorities of Italian semantic analysis of text society. From the basic necessities of home and rent to the complexities of the economy and politics, these words refer to some of the challenges and opportunities individuals and institutions face.

A new index of importance for economic keywords

Table 5 demonstrates the distribution of sentiment polarity of the extracted sentences across the four time periods by displaying the number and percentage of each sentence type in each period. Search results indicated that the first news article directly related to our study’s objective was published by The New York Times in 1980, and the current full year at the time of data collection was 2020. The authors wish to thank Vincenzo D’Innella Capano, CEO of Telpress International B.V., and to Lamberto Celommi, for making the news data available. The computing resources and the related technical support used for this work were provided by CRESCO/ENEAGRID High Performance Computing infrastructure and its staff.

LR and MNB are statistical models that make predictions by considering the probability of class based on a decision boundary and the frequency of words in sentences, respectively. Similarly, LR and SVC employed a boundary to predict the class using a features map of words. SGD served as an optimization method that enhanced classifier performance for SVC and LR models. RF utilized a boosting technique by combining multiple decision trees and making predictions based on the voting results from each tree. Following model construction, hyperparameters were fine-tuned using GridSearchCV. This method systematically searched for optimal hyperparameters within subsets of the hyperparameter space to achieve the best model performance.

A psycholinguistic study of intergroup bias and its cultural propagation

I’d like to express my deepest gratitude to Javad Hashemi for his constructive suggestions and helpful feedback on this project. Particularly, I am grateful for his insights on sentiment complexity and his optimized solution to calculate vector similarity between two lists of tokens that I used in the list_similarity function. As the classification report shows, the TopSSA model achieves better accuracy and F1 scores reaching as high as about 84%, a significant achievement for an unsupervised model. If the S3 is positive, we can classify the review as positive, and if it is negative, we can classify it as negative.

Additionally, this approach is inspired by the human brain and requires extensive training data and features, eliminating manual selection and allowing for efficient extraction of insights from large datasets23,24. The diverse opinions and emotions expressed in these comments are challenging to comprehend, as public opinion on war events can fluctuate rapidly due to public debates, official actions, or breaking news13. Managing hate speech and offensive remarks in war discussions on YouTube is crucial, requiring an understanding of user-generated content, privacy, and moral considerations, especially during wartime14,15. The unstructured nature of YouTube comments, the use of colloquial language, and the expression of a wide range of opinions and emotions present challenges for this task.

The final sample comprised over 1,808,000 news articles published between January 2, 2017, and August 30, 2020. Our textual analysis focused solely on the initial 30% of each news article, including the title and lead. This decision aligns with previous research21 and is based on the understanding that online news readers tend only to skim the beginning of an article, paying particular attention to the title and opening paragraphs43,44. As a robustness check, we ran our models on the full text of the articles but found no significant improvement in results. This scenario, simple though it may seem, shows how effectively sentiment analysis can improve customer outcomes.

As with the other forecasting models, we implemented an expanding window approach to generate our predictions. Specifically, we started with an initial subset of data to train the neural network and make a first prediction for the next period. The training set window was ChatGPT subsequently expanded by including the next observation, and the process was repeated recursively. Telpress International B.V.—a company that collects online news from multiple web sources, including mainstream media sites and blogs—provided access to online news data.

Modesty is highly valued in many Middle Eastern cultures to preserve honour and maintain social order (Ennaji and Sadiqi, 2011). Unwanted sexual attention is often seen as a violation of these cultural norms, leading to victim-blaming and shaming (Eltahawy, 2015). It is argued that the prevalence of unwanted sexual attention perpetuates a culture of fear and insecurity for women in the Middle East. It restricts their freedom of movement and limits their opportunities for education and employment, hindering their overall empowerment (Bouhlila, 2019). In cases of sexual coercion, victims often face immense pressure to remain silent due to fears that their reputation or family’s honour will be tarnished, which perpetuates a cycle of violence and oppression within Middle Eastern societies. Victims often find themselves trapped in abusive relationships without access to legal protection or support systems, leading to long-term psychological trauma.

The accuracy, precision, and recall of the Bi-LSTM for Amharic sentiment dataset were 85.27 percent, 85.24%, and 81.67%, respectively. The result shows that BI-LSTM model performs better than CNN model which further indicates the capability of BI-LSTM to improve the classification performance by considering the previous and future words during learning. The Dravidian Code-Mix-FIRE 2020 has been informed of the sentiment polarity of code-mixed languages like Tamil-English and Malayalam-English14.

Word2Vec was utilized for word embedding, combining Convolutional Neural Networks (CNN) with recurrent neural networks (RNN). Despite achieving 88.3% and 47.5% accuracy, the hybrid model was deemed suboptimal, suggesting further experimentation with different RNN models. The non-i.i.d learning paradigm of gradual machine learning (GML) was originally proposed for the task of entity resolution8.

The training accuracy increases as the number of epochs increases, but the Validation accuracy decreases as the number of epochs increases. When compared to the work required to combat over-fitting, building a model and executing the code is the easier part. The researcher used many regularization approaches for our model, such as Seeding (also known as Random state) from 42 to 50.

This adaptive mechanism allows LSTMs to discern the importance of data, enhancing their ability to retain crucial information for extended periods28. Chatbots help customers immensely as they facilitate shipping, answer queries, and also offer personalized guidance and input on how to proceed further. Moreover, some chatbots are equipped with emotional intelligence that recognizes the tone of the language and hidden sentiments, framing emotionally-relevant responses to them. Google incorporated ‘semantic analysis’ into its framework by developing its tool to understand and improve user searches.

LDA allows a set of news stories and tweets to be categorized into their underlying topics. According to Atkins et al. (2018) “a topic is a set of words, where each word has a probability of appearance in documents labeled with the topic. Each document is a mixture of corpus-wide topics, and each word is drawn from one of these topics. We have followed Atkins’ methodology to assess whether topics extracted from tweets and news headlines can be used to predict directional changes in market volatility.

These pre-trained models are trained on large corpus in order to capture long-term semantic dependencies. This feature refers to a sentiment analysis tool’s capability to analyze text in multiple languages. Multilingual support is essential in preventing biases, as it promotes an inclusive understanding of languages and cultures and ensures sentiment from global customers is recognized. Understanding multiple languages also helps in training models to understand the complexities of words, phrases, and slang, as one positive or negative sentiment might mean neutral in another language. Sentiment analysis tools determine the positive-negative polarity of user-generated text at their most basic level, and offer more advanced tools for working with larger datasets.

Let us now describe the steps we took to perform LDA and use the obtained topic distribution to predict next day’s market volatility (“UP” or “DOWN”). A growing number of research papers use Natural Language Processing (NLP) methods to analyze how sentiment of firm-specific news, financial reports, or social media impact stock market returns. You can foun additiona information about ai customer service and artificial intelligence and NLP. An important early work by Tetlock (2007) explores possible correlations between the media and the stock market using information from the Wall Street Journal and finds that high pessimism causes downward pressure on market prices.

Overfitting occurs when a model becomes too specialized in the training data and fails to generalize well to unseen data. To address these issues, it is recommended to increase the sample size by including more diverse and distinct samples in each class. A larger sample size helps to capture a wider range of patterns and reduces the risk of overfitting. Additionally, incorporating more varied samples can help mitigate the sensitivity caused by high-frequency words. Furthermore, it is important to consider the limitations of training models in a specific context, such as sexual harassment in Middle East countries. Models trained on such data may not perform as expected when applied to datasets from different contexts, such as anglophone literature from another region.

Also, LDA is a generative unsupervised statistical algorithm for extracting thematic information (topics) of a collection of documents within the Bayesian statistical paradigm. The LDA model assumes that each document is made up of various topics, where each topic is a probability distribution over words. A significant advantage of using the LDA model is that topics can be inferred from a given collection without input from any prior knowledge. To summarize the results obtained in this experiment, the results from CNN-Bi-LSTM achieved better results than those from the other Deep Learning as shown in the Fig.

With that said, sentiment analysis is highly complicated since it involves unstructured data and language variations. Sentiment analysis tools enable businesses to understand the most relevant and impactful feedback from their target audience, providing more actionable insights for decision-making. The best sentiment analysis tools go beyond the basics of positivity and negativity and allow users to recognize subtle emotions, more holistic contexts, and sentiment across diverse channels.

The output was then passed into the fully connected layer with Sigmoid as the binary classifier. Data preprocessing is the process of removing distortion from data to make any classification task easier in our case sentiment classification and improve the performance of the model. As a result, it is critical to apply data preprocessing to overcome such issues because the more the data is cleaned the more accurate the deep learning model will be. Each word is assigned a continuous vector that belongs to a low-dimensional vector space. Neural networks are commonly used for learning distributed representation of text, known as word embedding27,29.

In addition, any posts by users who posted more than one message or cross-posted in both conditions were removed. The final sample size consisted of 8690 messages (5703 from the depression forum condition). Our sample size consisted of 26,473,715 tweets, all were in the English language, and all were original (i.e., retweets were filtered out). Text cleaning included the removal of links, tags, and emoticons before any linguistic analysis. Tweets were collected via Twitter’s dedicated API from across the United States, including all 50 states and the District of Columbia.

It can be written connected or disconnected at the end, placed within the word, or found at the beginning. Besides, diacritics or short vowels control the word phonology and alter its meaning. These characteristics propose challenges to word embedding and representation21. Further challenges for Arabic language processing are dialects, morphology, orthography, phonology, and stemming21. In addition to the Arabic nature related challenges, the efficiency of word embedding is task-related and can be affected by the abundance of task-related words22. Therefore, a convenient Arabic text representation is required to manipulate these exceptional characteristics.

Firms and governments are looking for useful information in these user comments such as the feelings behind client comments17. SA refers to the application of machine and deep learning and computational linguistics to investigate ChatGPT App the feelings or views expressed in user-written comments18,19. Because of increasing interest in SA, businesses are interested in driving campaigns, having more clients, overcoming their weaknesses, and winning marketing tactics.

Azure AI Language translates more than 100 languages and dialects, including some deemed at-risk and endangered. These Internet buzzwords contain rich semantic and emotional information, but are difficult to be recognized by general-purpose lexical tools. Danmaku domain lexicon can effectively solve this problem by automatically recognizing and manually annotating these neologisms into the lexicon, which in turn improves the accuracy of downstream danmaku sentiment analysis task. Table 6 More pronounced are the effects observed from the removal of syntactic features and the MLEGCN and attention mechanisms.