<p>The Internet of Things (IoT) has transformed various aspects of daily life by generating vast, heterogeneous data, necessitating efficient data filtering to ensure optimal system performance. Recommender systems are crucial in delivering personalized recommendations based on user preferences and behavioral patterns. However, a significant challenge lies in effectively categorizing and analyzing large-scale, diverse, and complex data to extract meaningful insights for accurate recommendations. This research presents an innovative model that integrates ontology with clustering methods and deep neural networks, significantly enhancing the accuracy of IoT recommender systems by improving the data filtering process. Unlike traditional methods, this model has wide-ranging applications in optimizing services across diverse IoT recommender system environments. By leveraging semantic relationships and structured knowledge representations, our approach surpasses conventional filtering techniques, leading to more refined recommendations. The proposed model not only improves user experience but also optimizes service delivery in dynamic IoT environments. To assess its effectiveness, we conducted extensive experimental evaluations and benchmarked our model against existing state-of-the-art methods. The results demonstrate substantial improvements in recommendation accuracy and a notable reduction in error rates. Specifically, our approach achieves approximately a 50% improvement over traditional collaborative filtering methods, paving the way for more intelligent, adaptive, and personalized IoT-based recommender systems. </p>