Optimizing Lung Cancer Detection: The Synergy of Support Vector Machine and Random Forest

Lung cancer is still a major cancer killer globally and finding the disease at an early stage is crucial to success. The objective of this work is to improve the accuracy of the lung cancer detection system employing a combined model of machine learning algorithms. The study aims to compare the performance of chosen Support Vector Machine (SVM) and Random Forest (RF) classifiers called the SVRF model to enhance reliability and precision in classification. The dataset obtained from Kaggle is first pre-processed to involve cleaning of data, addressing class imbalance through the SMOTE algorithm, and managing outliers via z-score. Feature selection is done using Information Gain, by finding the most important indicators. The result of the proposed SVRF model has an average accuracy of 91.19% through experimentation, compared to prior work which only achieved the accuracy of up to 80% by using conventional machine learning and deep neural networks. The hybrid approach is efficient in balancing the measures of precision and recall, especially in identifying positive lung cancer cases. These results demonstrate how the proposed model can be used in clinical practice, and thus yields more accurate and precise diagnostic information. There is a future work, that will narrow down the presented work to enhance the recognition of negative samples, thus helping to early diagnose and treat lung cancer patients.