Description:FIELD OF DISCLOSURE
[0001] The present disclosure relates generally relates to the field of artificial intelligence and natural language processing. More specifically, it pertains to a hybrid sentiment analysis system for twitter integrating natural language processing and explainable AI.
BACKGROUND OF THE DISCLOSURE
[0002] The analysis of human emotions, opinions, and attitudes has long been a central area of inquiry in the fields of linguistics, psychology, and computer science. With the rise of digital communication platforms, sentiment analysis has evolved into one of the most prominent applications of natural language processing (NLP). Sentiment analysis refers to the automated computational study of subjective information such as opinions, evaluations, appraisals, attitudes, and emotions expressed in written or spoken language. The primary goal of sentiment analysis is to determine the polarity of a text whether it conveys positive, negative, or neutral sentiment. As digital interactions expand in volume and complexity, the demand for more accurate and transparent sentiment analysis systems has increased, particularly in platforms characterized by rapid and opinion-heavy exchanges such as Twitter.
[0003] Historically, sentiment analysis techniques were grounded in rule-based and lexicon-driven approaches. Early systems relied on sentiment dictionaries that classified words as positive, negative, or neutral. These methods provided a baseline for understanding textual emotions but were limited in handling contextual nuances, sarcasm, and domain-specific expressions. As Twitter became a prominent platform for public opinion, the shortcomings of lexicon-based systems became increasingly evident. The inability of such methods to process slang, acronyms, and evolving cultural references limited their applicability to social media contexts. Moreover, sentiment expressed in tweets often relies heavily on context, sarcasm, or even multimodal elements such as images and GIFs, making purely dictionary-based approaches inadequate.
[0004] The advent of machine learning introduced a paradigm shift in sentiment analysis. Traditional supervised learning methods such as Support Vector Machines (SVM), Naïve Bayes, and logistic regression were applied to classify sentiments in Twitter data. By training models on annotated datasets, these algorithms achieved greater accuracy than lexicon-based methods, particularly in recognizing patterns beyond simple word polarity. However, these approaches still faced limitations in managing the dynamic nature of Twitter language and the complexities introduced by irony and cultural variation. Additionally, feature engineering—where researchers manually crafted linguistic features such as n-grams, parts of speech tags, and syntactic dependencies was labor-intensive and often failed to generalize across diverse datasets.
[0005] The rise of deep learning further transformed sentiment analysis by reducing reliance on manual feature engineering. Neural network architectures such as Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) demonstrated their capacity to automatically learn hierarchical and sequential features from text data. RNN variants such as Long Short-Term Memory (LSTM) networks and Gated Recurrent Units (GRUs) were particularly effective at capturing long-range dependencies in text, enabling them to process complex sentence structures and contexts. With the growth of publicly available Twitter datasets, researchers successfully trained these deep learning models to outperform traditional machine learning methods in sentiment classification tasks.
[0006] More recently, the development of transformer-based architectures has revolutionized NLP, including sentiment analysis. Transformers, exemplified by models such as BERT (Bidirectional Encoder Representations from Transformers) and RoBERTa, leverage self-attention mechanisms to capture contextual relationships between words in a sentence with remarkable efficiency. Unlike RNNs, transformers process entire sequences in parallel, enabling them to capture both local and global dependencies simultaneously. This breakthrough has allowed for state-of-the-art performance in various NLP tasks, including sentiment classification on Twitter datasets. Pretrained transformer models fine-tuned on domain-specific Twitter corpora have become widely adopted due to their adaptability and superior accuracy.
[0007] Despite these advancements, sentiment analysis on Twitter continues to face significant challenges. One of the most prominent issues is the presence of noise and ambiguity in tweets. Tweets often contain misspellings, non-standard grammar, emojis, hashtags, and hyperlinks that complicate text preprocessing and interpretation. Furthermore, the multilingual nature of Twitter adds layers of complexity, as sentiment analysis systems must adapt to diverse languages and dialects. Sarcasm and irony remain particularly difficult to detect, as their interpretation often relies on extralinguistic knowledge or world context. For example, a tweet stating, “Great, another traffic jam!” may appear positive on the surface but is clearly negative when understood in context.
[0008] Another major limitation in sentiment analysis systems is the lack of transparency and explainability. Deep learning models, particularly those based on neural networks, are often regarded as “black boxes” due to their complex internal computations that are difficult for humans to interpret. While such models achieve high predictive accuracy, their opacity hinders trust and accountability, especially in applications where decisions may have regulatory or societal consequences. For instance, when analyzing public sentiment about political candidates, stock market trends, or pandemic policies, stakeholders often demand not only accurate predictions but also explanations for why a particular sentiment classification was made. This demand has spurred research into explainable AI (XAI) methods that aim to provide human-interpretable justifications for machine learning predictions.
[0009] Explainable AI techniques for sentiment analysis have developed along multiple dimensions. Post-hoc methods attempt to interpret existing models by highlighting the most influential words or features contributing to a sentiment decision. For example, visualization tools such as saliency maps or attention heatmaps identify which parts of a tweet influenced the classification most strongly. Intrinsic methods, on the other hand, aim to design inherently interpretable models that maintain transparency without the need for post-hoc analysis. Both approaches are gaining traction, but striking the right balance between accuracy and explainability remains a central challenge.
[0010] Beyond technical considerations, sentiment analysis on Twitter also intersects with broader societal and ethical issues. Public opinion mining has become a powerful tool for governments, corporations, and social organizations. Companies analyze tweets to gauge customer satisfaction, identify emerging trends, and track brand reputation. Political organizations monitor sentiment to understand voter attitudes and shape campaign strategies. Public health authorities analyze Twitter discussions to track disease outbreaks or measure sentiment toward vaccination campaigns. While these applications underscore the utility of sentiment analysis, they also raise concerns regarding privacy, bias, and misuse. For instance, biased training datasets can produce skewed sentiment predictions, amplify stereotypes or misrepresenting public opinion. Moreover, the potential for manipulation exists if sentiment analysis tools are used unethically, such as in microtargeted political advertising.
[0011] The evolution of sentiment analysis techniques has also mirrored technological advancements in computational infrastructure. Early lexicon-based systems could be executed on modest hardware, while machine learning methods demanded more powerful processors and larger datasets. Deep learning, and particularly transformer-based approaches, require significant computational resources, including graphics processing units (GPUs) and tensor processing units (TPUs). This increasing demand has raised questions about the scalability and sustainability of advanced sentiment analysis methods. At the same time, innovations in edge computing and optimized neural network architectures are enabling more efficient deployments, including real-time sentiment monitoring on platforms like Twitter.
[0012] In parallel with these developments, hybrid approaches to sentiment analysis have emerged, integrating multiple methodologies to achieve higher performance and robustness. Hybrid systems often combine lexicon-based sentiment scoring with machine learning classifiers or fuse rule-based linguistic features with deep learning outputs. The motivation behind such integration lies in leveraging the strengths of different approaches while compensating for their weaknesses. For instance, lexicon methods provide interpretability, while neural networks offer accuracy. By combining them, hybrid systems can achieve a balance between transparency and predictive power. Such approaches are particularly relevant in Twitter analysis, where accuracy must coexist with user trust and explainability.
[0013] Research into sentiment analysis on Twitter has also been enriched by the availability of benchmark datasets and evaluation frameworks. Collections such as the Sentiment140 dataset, the Stanford Twitter Sentiment dataset, and SemEval competitions have provided standardized resources for training and benchmarking models. These datasets, however, are not without limitations. They may reflect biases inherent in the time and context of data collection, limiting their generalizability across evolving linguistic patterns. Additionally, annotation quality can vary, with subjectivity influencing how sentiment labels are assigned. As Twitter discourse evolves with new slang, memes, and cultural references, there is a continuous need for updated datasets and adaptive learning strategies.
[0014] The increasing importance of transparency in AI has also sparked regulatory interest. Emerging legal frameworks in data protection, algorithmic accountability, and AI ethics emphasize the need for systems that are not only accurate but also interpretable and auditable. In the European Union, for instance, regulations such as the General Data Protection Regulation (GDPR) emphasize the right to explanation, which mandates that individuals affected by automated decision-making have access to understandable justifications. In the context of sentiment analysis on Twitter, such regulations reinforce the demand for systems that can explain why a tweet was classified in a particular way.
[0015] As AI-driven sentiment analysis matures, its integration with other modalities of communication is becoming increasingly relevant. Twitter users often combine text with images, GIFs, or videos, all of which contribute to sentiment expression. Multimodal sentiment analysis, which integrates visual and textual cues, represents a frontier in the field. However, multimodal analysis also amplifies the challenges of explainability, as systems must account for how different modalities interact to produce a sentiment judgment. This complexity underscores the importance of hybrid and transparent approaches that can handle diverse input sources while remaining interpretable to end users.
[0016] Thus, in light of the above-stated discussion, there exists a need for a hybrid sentiment analysis system for twitter integrating natural language processing and explainable AI.
SUMMARY OF THE DISCLOSURE
[0017] The following is a summary description of illustrative embodiments of the invention. It is provided as a preface to assist those skilled in the art to more rapidly assimilate the detailed design discussion which ensues and is not intended in any way to limit the scope of the claims which are appended hereto in order to particularly point out the invention.
[0018] According to illustrative embodiments, the present disclosure focuses on a hybrid sentiment analysis system for twitter integrating natural language processing and explainable AI which overcomes the above-mentioned disadvantages or provide the users with a useful or commercial choice.
[0019] An objective of the present disclosure is to create an efficient pipeline for preprocessing Twitter data, including noise reduction, tokenization, and contextual embedding, to optimize input for hybrid sentiment modeling.
[0020] Another objective of the present disclosure is to design and implement a hybrid sentiment analysis system that combines traditional Natural Language Processing (NLP) techniques with advanced machine learning models for analyzing Twitter data.
[0021] Another objective of the present disclosure is to improve the accuracy of sentiment prediction on tweets by addressing challenges such as ambiguity, sarcasm, slang, and evolving linguistic expressions common in social media communication.
[0022] Another objective of the present disclosure is to integrate Explainable AI (XAI) mechanisms into the sentiment analysis process, thereby enhancing transparency and interpretability of predictions for end users, researchers, and policymakers.
[0023] Another objective of the present disclosure is to develop a context-aware sentiment classification framework capable of capturing the nuanced meaning of short and informal text messages typical of Twitter posts.
[0024] Another objective of the present disclosure is to ensure that the sentiment analysis system adapts to dynamic and evolving language trends on social media, enabling long-term applicability and robustness.
[0025] Another objective of the present disclosure is to evaluate the performance of the hybrid system using benchmark datasets and real-time Twitter streams, measuring accuracy, precision, recall, F1-score, and explainability.
[0026] Another objective of the present disclosure is to enhance user trust in sentiment analysis outputs by providing interpretable reasoning behind classification decisions, reducing the “black-box” effect of deep learning models.
[0027] Another objective of the present disclosure is to explore the potential of the hybrid system for real-world applications, such as trend monitoring, political discourse analysis, brand reputation management, and crisis detection.
[0028] Yet another objective of the present disclosure is to optimize the system for scalability and real-time performance, enabling deployment in large-scale Twitter sentiment monitoring scenarios.
[0029] In light of the above, a hybrid sentiment analysis system for twitter integrating natural language processing and explainable AI comprises a data acquisition module configured to collect tweets from social media platforms via application programming interfaces (APIs). The system also includes a preprocessing and feature extraction module configured to clean the collected tweets by removing symbols, stopwords, and noise. The system also includes a sentiment classification module configured to classify sentiment of the processed tweets. The system also includes an explainable AI (XAI) module configured to identify influential words or phrases contributing to sentiment classification and generate human-interpretable visual explanations. The system also includes a real-time visualization module configured to display sentiment predictions and corresponding model explanations on a dashboard for enhanced transparency and interpretability. The system also includes an evaluation module configured to assess prediction quality using metrics, and to evaluate interpretability using feature importance ranking or XAI-specific metrics.
[0030] In one embodiment, the preprocessing and feature extraction module employs natural language processing techniques including word embeddings, BERT-based contextual representations, and syntactic structure analysis to enhance semantic understanding of tweets.
[0031] In one embodiment, the sentiment classification module combines rule-based NLP methods for detecting explicit sentiment words with machine learning models comprising long short-term memory (LSTM) networks, Transformer architectures, or fine-tuned BERT models to capture deep linguistic patterns and contextual nuances.
[0032] In one embodiment, the sentiment classification module is fine-tuned specifically for short-text inputs like tweets to maintain context-sensitive sentiment detection.
[0033] In one embodiment, the explainable AI module provides interactive visualizations highlighting the contribution of specific words or features to the sentiment decision, enabling users to trace and validate the reasoning behind predictions.
[0034] In one embodiment, the explainable AI module utilizes tools selected from SHAP or LIME to highlight influential words or phrases that contribute to sentiment classification and generates visual explanations to provide transparency of the model’s decision-making.
[0035] In one embodiment, the real-time visualization module presents sentiment predictions along with corresponding model explanations on an interactive dashboard to enable user interpretation, monitoring, and trust in predictive outcomes.
[0036] In one embodiment, the evaluation module measures prediction accuracy using metrics including F1-score, precision, and recall, and assesses interpretability using feature importance ranking and XAI-specific evaluation metrics.
[0037] In one embodiment, the data acquisition module continuously collects streaming tweets in real-time and stores them in a structured format for subsequent preprocessing, feature extraction, and sentiment analysis.
[0038] In one embodiment, the system is configured to detect and handle challenges in social media text including sarcasm, ambiguity, abbreviations, and evolving language patterns to improve classification accuracy.
[0039] These and other advantages will be apparent from the present application of the embodiments described herein.
[0040] The preceding is a simplified summary to provide an understanding of some embodiments of the present invention. This summary is neither an extensive nor exhaustive overview of the present invention and its various embodiments. The summary presents selected concepts of the embodiments of the present invention in a simplified form as an introduction to the more detailed description presented below. As will be appreciated, other embodiments of the present invention are possible utilizing, alone or in combination, one or more of the features set forth above or described in detail below.
[0041] These elements, together with the other aspects of the present disclosure and various features are pointed out with particularity in the claims annexed hereto and form a part of the present disclosure. For a better understanding of the present disclosure, its operating advantages, and the specified object attained by its uses, reference should be made to the accompanying drawings and descriptive matter in which there are illustrated exemplary embodiments of the present disclosure.
BRIEF DESCRIPTION OF THE DRAWINGS
[0042] To describe the technical solutions in the embodiments of the present disclosure or in the prior art more clearly, the following briefly describes the accompanying drawings required for describing the embodiments or the prior art. Apparently, the accompanying drawings in the following description merely show some embodiments of the present disclosure, and a person of ordinary skill in the art can derive other implementations from these accompanying drawings without creative efforts. All of the embodiments or the implementations shall fall within the protection scope of the present disclosure.
[0043] The advantages and features of the present disclosure will become better understood with reference to the following detailed description taken in conjunction with the accompanying drawing, in which:
[0044] FIG. 1 illustrates a flowchart outlining sequential step involved in a hybrid sentiment analysis system for twitter integrating natural language processing and explainable AI, in accordance with an exemplary embodiment of the present disclosure;
[0045] FIG. 2 illustrates the architectural flow diagram of a hybrid sentiment analysis system for twitter integrating natural language processing and explainable AI, in accordance with an exemplary embodiment of the present disclosure.
[0046] Like reference, numerals refer to like parts throughout the description of several views of the drawing;
[0047] The hybrid sentiment analysis system for twitter integrating natural language processing and explainable AI, which like reference letters indicate corresponding parts in the various figures. It should be noted that the accompanying figure is intended to present illustrations of exemplary embodiments of the present disclosure. This figure is not intended to limit the scope of the present disclosure. It should also be noted that the accompanying figure is not necessarily drawn to scale.
DETAILED DESCRIPTION OF THE DISCLOSURE
[0048] The following is a detailed description of embodiments of the disclosure depicted in the accompanying drawings. The embodiments are in such detail as to communicate the disclosure. However, the amount of detail offered is not intended to limit the anticipated variations of embodiments; on the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the present disclosure.
[0049] In the following description, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the present disclosure. It may be apparent to one skilled in the art that embodiments of the present disclosure may be practiced without some of these specific details.
[0050] Various terms as used herein are shown below. To the extent a term is used, it should be given the broadest definition persons in the pertinent art have given that term as reflected in printed publications and issued patents at the time of filing.
[0051] The terms “a” and “an” herein do not denote a limitation of quantity, but rather denote the presence of at least one of the referenced items.
[0052] The terms “having”, “comprising”, “including”, and variations thereof signify the presence of a component.
[0053] Referring now to FIG. 1 to FIG. 2 to describe various exemplary embodiments of the present disclosure. FIG. 1 illustrates a flowchart outlining sequential step involved in a hybrid sentiment analysis system for twitter integrating natural language processing and explainable AI, in accordance with an exemplary embodiment of the present disclosure.
[0054] A hybrid sentiment analysis system for twitter integrating natural language processing and explainable AI 100 comprises a data acquisition module 102 configured to collect tweets from social media platforms via application programming interfaces (APIs). The data acquisition module 102 continuously collects streaming tweets in real-time and stores them in a structured format for subsequent preprocessing, feature extraction, and sentiment analysis.
[0055] The system also includes a preprocessing and feature extraction module 104 configured to clean the collected tweets by removing symbols, stopwords, and noise. The preprocessing and feature extraction module 104 employs natural language processing techniques including word embeddings, BERT-based contextual representations, and syntactic structure analysis to enhance semantic understanding of tweets.
[0056] The system also includes a sentiment classification module 106 configured to classify sentiment of the processed tweets. The sentiment classification module 106 combines rule-based NLP methods for detecting explicit sentiment words with machine learning models comprising long short-term memory (LSTM) networks, Transformer architectures, or fine-tuned BERT models to capture deep linguistic patterns and contextual nuances. The sentiment classification module 106 is fine-tuned specifically for short-text inputs like tweets to maintain context-sensitive sentiment detection.
[0057] The system also includes an explainable AI (XAI) module 108 configured to identify influential words or phrases contributing to sentiment classification and generate human-interpretable visual explanations. The explainable AI module 108 provides interactive visualizations highlighting the contribution of specific words or features to the sentiment decision, enabling users to trace and validate the reasoning behind predictions. The explainable AI module 108 utilizes tools selected from SHAP or LIME to highlight influential words or phrases that contribute to sentiment classification and generates visual explanations to provide transparency of the model’s decision-making.
[0058] The system also includes a real-time visualization module 110 configured to display sentiment predictions and corresponding model explanations on a dashboard for enhanced transparency and interpretability. The real-time visualization module 110 presents sentiment predictions along with corresponding model explanations on an interactive dashboard to enable user interpretation, monitoring, and trust in predictive outcomes.
[0059] The system also includes an evaluation module 112 configured to assess prediction quality using metrics, and to evaluate interpretability using feature importance ranking or XAI-specific metrics. The evaluation module 112 measures prediction accuracy using metrics including F1-score, precision, and recall, and assesses interpretability using feature importance ranking and XAI-specific evaluation metrics.
[0060] The system is configured to detect and handle challenges in social media text including sarcasm, ambiguity, abbreviations, and evolving language patterns to improve classification accuracy.
[0061] FIG. 1 illustrates a flowchart outlining sequential step involved in a hybrid sentiment analysis system for twitter integrating natural language processing and explainable AI.
[0062] At 102, the process begins with the data acquisition module, which is responsible for collecting real-time tweets from social media platforms through application programming interfaces (APIs). This module continuously streams textual data, ensuring that the system receives up-to-date information reflecting the dynamic nature of social discourse on platforms like Twitter. By automating the collection of tweets, the system establishes a robust foundation for sentiment analysis, providing a diverse and representative dataset for downstream processing.
[0063] At 104, once tweets are acquired, they are passed to the preprocessing and feature extraction module, where the raw textual data undergoes extensive cleaning to remove noise such as symbols, irrelevant characters, URLs, and stopwords. This cleaning process ensures that only meaningful linguistic content is preserved, reducing the likelihood of erroneous sentiment detection due to irrelevant information. The module also performs feature extraction using advanced NLP techniques, generating word embeddings and syntactic structures that capture both semantic meaning and grammatical relationships within the tweets. By converting textual data into structured representations, this module allows the system to leverage the deep learning capabilities of subsequent classification models while maintaining sensitivity to the nuanced patterns in natural language.
[0064] At 106, the processed and feature-rich data is then transmitted to the sentiment classification module. This module implements a hybrid approach that combines rule-based NLP techniques with advanced machine learning models, including long short-term memory (LSTM) networks, Transformer architectures, and BERT-based models. The rule-based component detects explicit sentiment expressions such as positive or negative keywords, while the machine learning models identify more complex, context-dependent patterns in the text, including idiomatic expressions, sarcasm, or subtle tonal variations. By integrating these complementary techniques, the classification module produces accurate sentiment predictions that reflect both surface-level linguistic cues and deeper semantic structures.
[0065] At 108, following sentiment classification, the explainable AI module processes the outputs to enhance transparency and interpretability. The XAI module applies methods such as SHAP and LIME to highlight the most influential words or phrases that contributed to each sentiment decision. This generates human-interpretable visual explanations, allowing users to understand why a tweet was classified as positive, negative, or neutral. By making the decision-making process of the hybrid model visible, the XAI module addresses the black box nature of deep learning, providing a level of trust and accountability that is critical for applications such as brand monitoring, political trend analysis, or customer feedback evaluation.
[0066] At 110, the real-time visualization module then takes the sentiment predictions and explanatory outputs and presents them on an interactive dashboard. This interface enables users to monitor trends, identify patterns, and analyze sentiment dynamics efficiently. The dashboard highlights not only the classification results but also the underlying reasoning provided by the XAI module, ensuring that stakeholders can interpret and act upon the insights with confidence. The visualizations make it easier to identify spikes in positive or negative sentiment, track topic-specific discussions, and understand the textual features driving these trends.
[0067] At 112, the evaluation module continuously monitors the performance of the system. Prediction quality is assessed using metrics such as accuracy, F1-score, and other standard classification indicators, while interpretability is evaluated through feature importance ranking and XAI-specific measures. This module ensures that the hybrid system maintains high predictive reliability while preserving transparency, allowing for iterative improvements to both the NLP models and the explainability framework. The closed-loop integration of data acquisition, preprocessing, hybrid classification, explainable AI, visualization, and evaluation enables the system to provide timely, accurate, and interpretable sentiment analysis, making it a powerful tool for extracting meaningful insights from the ever-changing landscape of social media content.
[0068] FIG. 2 illustrates the architectural flow diagram of a hybrid sentiment analysis system for twitter integrating natural language processing and explainable AI.
[0069] the framework begins with preprocessing and feature extraction. This stage involves preparing raw data, such as text, speech, or other forms of unstructured input, so that it becomes suitable for analysis. Noise reduction, tokenization, stemming, and feature engineering are commonly performed here to highlight the most important aspects of the data. The aim of this step is to transform raw information into meaningful numerical or symbolic features that can be efficiently processed by machine learning algorithms.
[0070] Once the features are extracted, the process moves to the sentiment classification stage. Here, machine learning or deep learning models classify the data into categories such as positive, negative, or neutral sentiments. This stage is central to the system because it provides actionable insights based on the tone, emotion, or attitude expressed in the input. However, one of the biggest challenges in sentiment classification is that models often operate like black boxes, making it difficult to understand how or why a decision was made. This is where the need for transparency arises.
[0071] To address the challenge of hidden decision-making, the interpretability and transparency module is introduced. This part of the framework ensures that the system does not simply produce results but also explains the reasoning behind them. By incorporating interpretability, the AI system can show which features or inputs influenced a sentiment classification outcome. This not only builds trust among end-users but also supports developers, regulators, and researchers in understanding the underlying decision processes of the AI system. Transparency ensures that the system remains accountable and avoids biases that may otherwise go unnoticed.
[0072] At the center of this framework is Explainable AI, which integrates inputs from both the sentiment classification and interpretability layers. Explainable AI acts as the unifying element, combining classification outputs with explanations to deliver both predictions and justifications. For instance, if a review is classified as negative, Explainable AI clarifies whether specific keywords, sentence structures, or tone patterns influenced that outcome. This integration enables end-users to move beyond accepting AI decisions blindly, allowing them to evaluate and validate the reasoning behind model outputs.
[0073] While the invention has been described in connection with what is presently considered to be the most practical and various embodiments, it will be understood that the invention is not to be limited to the disclosed embodiments, but on the contrary, is intended to cover various modifications and equivalent arrangements included within the scope of the appended claims.
[0074] A person of ordinary skill in the art may be aware that, in combination with the examples described in the embodiments disclosed in this specification, units and algorithm steps may be implemented by electronic hardware, computer software, or a combination thereof.
[0075] The foregoing descriptions of specific embodiments of the present disclosure have been presented for purposes of illustration and description. They are not intended to be exhaustive or to limit the present disclosure to the precise forms disclosed, and many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described to best explain the principles of the present disclosure and its practical application, and to thereby enable others skilled in the art to best utilize the present disclosure and various embodiments with various modifications as are suited to the particular use contemplated. It is understood that various omissions and substitutions of equivalents are contemplated as circumstances may suggest or render expedient, but such omissions and substitutions are intended to cover the application or implementation without departing from the scope of the present disclosure.
[0076] Disjunctive language such as the phrase “at least one of X, Y, Z,” unless specifically stated otherwise, is otherwise understood with the context as used in general to present that an item, term, etc., may be either X, Y, or Z, or any combination thereof (e.g., X, Y, and/or Z). Thus, such disjunctive language is not generally intended to, and should not, imply that certain embodiments require at least one of X, at least one of Y, or at least one of Z to each be present.
[0077] In a case that no conflict occurs, the embodiments in the present disclosure and the features in the embodiments may be mutually combined. The foregoing descriptions are merely specific implementations of the present disclosure, but are not intended to limit the protection scope of the present disclosure. Any variation or replacement readily figured out by a person skilled in the art within the technical scope disclosed in the present disclosure shall fall within the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure shall be subject to the protection scope of the claims.
, Claims:I/We Claim:
1. A hybrid sentiment analysis system for twitter integrating natural language processing and explainable AI (100) comprising:
a data acquisition module (102) configured to collect tweets from social media platforms via application programming interfaces (APIs);
a preprocessing and feature extraction module (104) configured to clean the collected tweets by removing symbols, stopwords, and noise;
a sentiment classification module (106) configured to classify sentiment of the processed tweets;
an explainable AI (XAI) module (108) configured to identify influential words or phrases contributing to sentiment classification and generate human-interpretable visual explanations;
a real-time visualization module (110) configured to display sentiment predictions and corresponding model explanations on a dashboard for enhanced transparency and interpretability;
an evaluation module (112) configured to assess prediction quality using metrics, and to evaluate interpretability using feature importance ranking or XAI-specific metrics.
2. The system (100) as claimed in claim 1, wherein the preprocessing and feature extraction module (104) employs natural language processing techniques including word embeddings, BERT-based contextual representations, and syntactic structure analysis to enhance semantic understanding of tweets.
3. The system (100) as claimed in claim 1, wherein the sentiment classification module (106) combines rule-based NLP methods for detecting explicit sentiment words with machine learning models comprising long short-term memory (LSTM) networks, Transformer architectures, or fine-tuned BERT models to capture deep linguistic patterns and contextual nuances.
4. The system (100) as claimed in claim 1, wherein the sentiment classification module (106) is fine-tuned specifically for short-text inputs like tweets to maintain context-sensitive sentiment detection.
5. The system (100) as claimed in claim 1, wherein the explainable AI module (108) provides interactive visualizations highlighting the contribution of specific words or features to the sentiment decision, enabling users to trace and validate the reasoning behind predictions.
6. The system (100) as claimed in claim 1, wherein the explainable AI module (108) utilizes tools selected from SHAP or LIME to highlight influential words or phrases that contribute to sentiment classification and generates visual explanations to provide transparency of the model’s decision-making.
7. The system (100) as claimed in claim 1, wherein the real-time visualization module (110) presents sentiment predictions along with corresponding model explanations on an interactive dashboard to enable user interpretation, monitoring, and trust in predictive outcomes.
8. The system (100) as claimed in claim 1, wherein the evaluation module (112) measures prediction accuracy using metrics including F1-score, precision, and recall, and assesses interpretability using feature importance ranking and XAI-specific evaluation metrics.
9. The system (100) as claimed in claim 1, wherein the data acquisition module (102) continuously collects streaming tweets in real-time and stores them in a structured format for subsequent preprocessing, feature extraction, and sentiment analysis.
10. The system (100) as claimed in claim 1, wherein the system is configured to detect and handle challenges in social media text including sarcasm, ambiguity, abbreviations, and evolving language patterns to improve classification accuracy.