Aim: This work offers and compares two machine learning strategies for developing an efficient intrusion detection system, namely enhanced Novel Decision Tree and Linear Regression. Materials and Methods: Utilizing Intrusion Detection in conjunction with Novel Decision Trees and Linear Regression allows for the identification of network intrusions. With 20,000 records used for training and 8,000 records for testing, the total number of records utilized in this study is 28,000. By adjusting the Gpower settings to 0.05 and 0.85, the test achieves an average Gpower of almost 85%. Result: The two methods that were investigated for intrusion detection have a statistically significant difference, as shown by a significance value of 0.001 (p<0.05). When comparing accuracy, Novel Decision Tree (94.90%) performs better than Linear Regression (93.72%). The Linear Regression Classifier is among the most popular and easy-to-understand classification algorithms used today. The system's incapacity to react or halt assaults upon discovery is one of its flaws. Conclusion: The novel decision tree is more accurate than Linear Regression.