Prediction and Rule Generation for Leukemia using Decision Tree and Association Rule Mining

摘要

Leukemia is familiar as the cancer of the bone marrow. Bone marrow is known as the spongy tissue which creates blood cells. That’s why for blood cell productivity problems, Leukemia can boost heavily. It generally impacts the leukocytes or white blood cells. And also it grows gradually in human bodies. Both children and adults can be affected by leukemia. In this research, our objective is to detect leukemia applying Machine Learning algorithms, which can help medical science to detect leukemia. Therefore, we need a system to decide to detect Leukemia. That’s why we propose some well-known algorithms to compare accuracy to detect Leukemia in our study. We have used some different types of machine learning algorithms like Decision Tree, Logistic Regression and Random Forest algorithm to get the result that a patient is a cancer patient or not. We collected our data from Z H Shikder Medical College and Hospital, which includes 401 data in our dataset. Among the algorithms, the Decision tree algorithm gives the best result with 84.15% accuracy of prediction. In this paper, we show the comparative results analyzing different algorithms precision, recall and F1 score for our all data samples. We also try to extract rules from features with the help of Apriori algorithm and association rule.