ABSTRACT
Cancer is the most frequent illness that cannot be overlooked and ends in death; early diagnosis helps to navigate the most effective care to save human lives. Cancer can develop anywhere in the human body, but Hematologic (blood) cancer and solid tumor cancers are two main types of cancer. A solid tumor is a type of nodule, but not all types of nodules are cancer. In some instances, a cancer diagnosis is created on the doctor’s intuition, leading to the certain patient getting ignored and having complications. During the past few years, machine learning has been proved a popular and powerful method in many medical diagnosis areas and outperformed classical methods. This chapter mainly focuses on two types of nodules: Malignant (nodules that are cancer) and Benign (nodules that are not cancer). In this chapter, five supervised ML algorithms (Logistic Regression, SVM, Naïve Bayes, Decision Tree and KNN) are applied on two different types of datasets (breast and lung cancer). For cancer dataset classification main features of the algorithms are discussed. Their performance is analyzed based on their accuracy. The existing technique consists of five stages: pre-processing of the dataset, feature selection, training, testing the model and performance analysis.
