Category-Based Sentiment Analysis of Sindhi News Headlines Using Machine Learning, Deep Learning, and Transformer Models

The increasing reliance on digital journalism in regional languages has created a growing need for efficient and automated sentiment analysis systems to interpret public opinion. This study presents a web-based framework for Category-Based Sentiment Analysis of Sindhi News Headlines, focusing exclusively on a Deep Learning approach using a Recurrent Neural Network (RNN). The system enables users to register, upload datasets, and analyze textual data by classifying sentiments into Positive, Neutral, and Negative categories. Given the sequential and contextual nature of language, the RNN model is particularly well-suited for this task, as it effectively captures word dependencies and contextual relationships within sentences. The framework incorporates a pre-trained tokenizer and sequence padding techniques to ensure consistent and structured preprocessing of textual inputs.

The proposed system demonstrates high accuracy and reliability, with the RNN model achieving an accuracy of 98.87% in sentiment classification. By processing input text step-by-step, the model is able to understand subtle linguistic patterns and deliver precise predictions. In addition to classification, the system provides meaningful, sentiment-based suggestions that help users interpret the results effectively. This makes the framework valuable for journalists, researchers, and policymakers who seek insights from regional language content. By focusing solely on the RNN approach, the study emphasizes the strength of deep learning techniques in handling low-resource languages like Sindhi, showcasing their potential in advancing natural language processing applications and improving automated public opinion analysis.

Keywords

Sentiment Analysis, Sindhi News Headlines Recurrent Neural Network (RNN), Machine Learning, Deep Learning, Transformer Models, Natural Language Processing, Text Classification, Public Opinion Mining