DESIGN AND DEVELOPMENT OF ML MODEL BASED ON FEATURE EXTRACTION FOR CLASSIFICATION AND SURVIVAL PREDICTION IN LUNG CANCER PATIENTS

Abstract

Cancer is a multifactorial complicated illness that is caused by numerous gene mutations or dysregulation of gene interactions. This study, on the other hand, proposes a unique method to cancer categorization. A large dataset of tumors has been taken and use the Naive Bayes classifier to categorize them. The Class Topper Optimization method is used to extract the features. The CTO algorithm is a novel artificial intelligence technology that is rapidly converging. The suggested method is straightforward, less complicated, accurate, and has a low error rate, which is critical in cancer classification. The accuracy, precision, error rate, and classification efficacy of the suggested method are shown. The results are also compared to the KNN classifier, which has been used by numerous researchers in the past to classify cancer. Experiments on a variety of datasets revealed that our innovative technique was both reliable and effective. The results demonstrate that the proposed method is both quick and accurate, making it an excellent choice for real-world cancer diagnosis. Cancer has long been a major danger to human health and well-being, having posed the greatest challenge in the history of human illness. Cancer's high death rate is primarily owing to the disease's complexity and the wide range of clinical outcomes. Because cancer survival prediction has become a major focus of cancer research, it will be important to increase the accuracy of this prediction. Many models have been suggested at the moment, however most of them simply use single genetic data or clinical data to construct prediction models for cancer survival. There is a lot of emphasis in present survival studies on determining whether or not a patient will survive five years. The personal issue of how long a lung cancer patient will survive remains unanswered. The goal of this research is to estimate the overall survival time with lung cancer. Two machine learning challenges are derived from a single customized query. To begin with, determining whether a patient will survive for more than five years is a simple binary question. The second step is to develop a five-year survival model using regression analysis. When asked to forecast how long a lung cancer patient would survive within five years, our models are accurate to within a month of the mean absolute error (MAE). Lung tumors have been linked to a number of biomarker genes. As a part of our investigation into lung cancer prognosis, we integrated a feature selection method with a classification system. Using feature selection approaches to minimize the number of features, we believe that most classification systems may be improved. There are certain factors that have a greater impact on the categorization algorithms than others. The findings of our tests using a well-known classification technique, namely naive Bayes, SSA, have been provided. As a result, naïve Bayes provided superior output without SSA, but SSA enhanced performance. New algorithms and feature selection strategies will be tested in the future as part of this research. Our experiments will include both cluster and ensemble methods. Keywords: Lung Cancer, Machine Learning, Cancer prediction, Electronic nose, Accuracy, Cancer classification, Cancer detection, Heuristic Class Topper Optimization (HCTO), Naïve Bayes, precision, Recall, Squirrel search algorithm.