Explainable Sentiment Analysis for Social Media Post Using Attention Mechanisms

The project was entitled ” Explainable Sentiment Analysis for Social Media Posts Using Attention Mechanisms ” the exponential growth of social media platforms has led to a surge in user generated content rich with opinions and sentiments, making sentiment analysis, a key aspect of natural language processing (NLP), essential for classifying sentiments into positive, negative, or neutral categories. This process has critical applications in marketing, customer feedback analysis, and political sentiment tracking. However, the unstructured nature of social media text, coupled with challenges like noise, misspellings, and irrelevant content, necessitates robust preprocessing and advanced analytical methods. This study leverages attention mechanisms to enhance sentiment analysis by improving prediction accuracy and interpretability. Attention mechanisms enable models to focus on the most relevant text elements, making decision-making more transparent. Using a dataset of millions of pre labeled tweets, the system integrates machine learning techniques like Logistic Regression and Multilayer Perceptron with Transformer-based architectures to achieve high classification accuracy. Dimensionality reduction through Principal Component Analysis (PCA) ensures computational efficiency without compromising performance