Sentiment analysis has emerged as an essential technique for understanding public opinion in digital communication platforms. Government and organizational e-consultation modules enable citizens and users to submit feedback, suggestions, and complaints online. However, the large volume of textual comments generated through these platforms makes manual analysis difficult and time-consuming. This study proposes a sentiment analysis system for automatically classifying comments received through an e-consultation module into positive, negative, and neutral categories using machine learning and natural language processing techniques. The system utilizes text preprocessing methods such as tokenization, stop-word removal, stemming, and vectorization to prepare textual data for classification. Machine learning models are implemented using Python-based frameworks including TensorFlow, Keras, and Natural Language Toolkit (NLTK). The backend of the system is developed using Flask, while React is used for the user interface and MySQL for storing user comments and processed results. The proposed system enables administrators and decision-makers to quickly interpret public sentiment, identify major concerns, and prioritize issues based on collective feedback. The architecture incorporates a modular pipeline consisting of data acquisition, preprocessing, sentiment classification, and visualization of results. Experimental evaluation demonstrates that automated sentiment analysis significantly improves efficiency in processing large volumes of user feedback compared to manual evaluation. The system also supports realtime monitoring of public opinion, thereby enhancing transparency and responsiveness in digital governance systems. Overall, the proposed approach demonstrates how machine learning and NLP technologies can be effectively integrated with e-consultation platforms to improve policy analysis, decision-making, and citizen engagement
IJERST (Sat,) studied this question.