The rise in IoT devices has caused significant security concerns, necessitating robust IDS tailored for IoT environments. This article presents a novel approach that combines Supervised, Unsupervised, Semi-Supervised, and Reinforcement Learning techniques to develop an adaptable IDS for IoT networks. The proposed system uses different learning methods to analyze network traffic, detect malicious activities, and adapt to evolving threats. Eight different models are used for preliminary classification tasks in supervised learning models. Isolation Forest and K-Means are utilized in unsupervised learning for anomaly detection purposes. Semi-Supervised learning enhances model effectiveness by utilizing both labeled and unlabeled data, while Reinforcement Learning adapts the IDS to emerging threat patterns. Comprehensive evaluations conducted on a real IoT network traffic dataset demonstrate the success of proposed approach, achieving a 100% accuracy for most supervised learning models, as well as being capable of identifying 775 and 4629 anomalies, highlighting its ability to significantly improve the detection of malicious activity in IoT networks. Overall, the suggested technique enhances the ability of IDS in IoT networks to detect and adapt to security challenges, effectively addressing the evolving threats.