IJIRST –International Journal for Innovative Research in Science & Technology| Volume 4 | Issue 9 | February 2018 ISSN (online): 2349-6010 All rights reserved by www.ijirst.org 10 Comparing Existing Methods for Predicting the Detection of Possibilities of Blood Cancer by Analyzing Health Data Gandhi Priyank Sanjay Dr. Viral Nagori School of Doctoral Research & Innovations, GLS University School of Doctoral Research & Innovations, GLS University Abstract E-health is the word of this era and had been much popularized due to its advantages, but still doesn't have the clear definition of the word. Till 90's it was merely a word in a dictionary but now is often used in general practice and its more than just "internet medicine", it’s all about virtually relating computer, medicines, technology and healthcare. This term was previous used by industrialists and people related to marketing and business, but now it’s often used in academics. Machine Learning methods have change the way to deal with the diseases. For cancer, it may improve an understanding, the way how it progressed, the methods to validate and various ways to improve the health by minimizing its disastrous effect of the disease and improve the patients' health. The aim of the current research is to analyze large data obtained from health records using data mining and machine learning algorithms. The study of existing work is to understand better the current scenario and the work in medical science for detecting the blood cancer. Keywords: Blood Cancer, Mining, Cancer Detection, Health, ANN, Ehealth _______________________________________________________________________________________________________ I. INTRODUCTION E-health is the word of this era and and had been much popularized due to its advantages, but still doesn't have the clear definition of the word. Till 90's it was merely a word in a dictionary but now is often used in general practice and its more than just "internet medicine", it’s all about virtually relating computer, medicines, technology and healthcare. This term was previous used by industrialists and people related to marketing and business, but now it’s often used in academics.[1 8] Internet and its related technologies gaps the bridge in providing various medical information, health related services and detailed informatics of the outcome. This word, E-health is not just only a word but a technical revolution along with a change in mindset to think, connect and an attitude to commit for networked technologies, thinking globally and improving healthcare worldwide by availing and refining the availed information. [18] The second leading cause of death in the United States is Cancer or malignant neoplasm. Cancer consists of more than 100 different disease that may affect all major organs and every functioning of the body. [21] All cancers arise from abnormal and uncontrolled cell growth brought about by changes or damage in a cell's DNA. When mutations occur or a cell gains or loses a chromosome during mitosis, the biological pathways that normally inhibit cell division are disabled, resulting in a proliferation of damaged cells that may ultimately metastasize to other parts of the body. [3] Machine Learning methods have change the way to deal with the diseases. For cancer, it may improve an understanding, the way how it progressed, the methods to validate and various ways to improve the health by minimizing its disastrous effect of the disease and improve the patients' health. II. VARIOUS METHODS OF ML During the data preprocessing different types machine learning techniques are define such as 1) ANNs 2) DTs 3) SVMs 4) BNs is available ANN Variety of pattern recognition and problems related to classification are handled by ANN's. Combination of various input variables generate an output via a trained mechanism.[18] For this process multiple hidden layers that are interconnected typically known as neural connections are used. However, some drawbacks are also associated with ANN as the layered generic structure of ANN is time consuming resulting poor performance. This technique is often known as "black-box" technology. Trying to find out how classification is done as well as the reason for failure of ANN is almost impossible to detect. [18]