Recommender systems play a critical role in personalizing user experiences by suggesting relevant content based on preferences. Traditional collaborative filtering methods often struggle with challenges such as data sparsity and cold-start issues, limiting their effectiveness. This research proposes an enhanced collaborative recommender system that integrates Natural Language Processing (NLP) techniques to analyze textual reviews and extract latent features, such as sentiment and topics, to improve recommendation accuracy. The system is implemented using the RapidMiner tool, leveraging its user-friendly workflows for data preprocessing, feature extraction, and model training. Experimental evaluations conducted on the IMDb movie review dataset demonstrate that incorporating NLP features significantly improves recommendation performance, achieving a 15% enhancement in precision and reducing prediction errors compared to baseline models. This study highlights the potential of combining textual insights with traditional collaborative filtering methods, offering a practical approach to addressing data sparsity and cold-start challenges.