AI-Driven Deep Learning Based Sentiment Analysis with Optimization Healthcare Recommendation System Using CNN and biLSTM

Healthcare investigates sentimental analysis by analyzing patients' comments to identify their issues and recommend appropriate specialists. This datasheet contains 6000 patient comments, which were preprocessed using normalization, tokenization, label encoding, and sequence padding. Three models were applied here: Naive Bayes, Support vector machine (SVM), and CNN combined with the biLSM model, achieving an accuracy of 0.9828, 0.9552, and 0.9930, respectively. The hybrid model successfully captured low-level and high-level features by providing strong sentiment classification. This method overcomes problems like delaying diagnosis and constricted access to professionals by helping healthcare providers, which gives them insights into patients’ issues and needs for recommending accurate specialists. By simplifying the processing of patient feedback, it showcases how machine learning and deep learning approach problems and provide suitable solutions. For more trustworthiness and integrity in medical sectors, future research has the potential to improve the system by applying models and growing datasheets.