A Survey Paper on Evolving Techniques for the Prediction of Type 2 Diabetes. RATNA NITIN PATIL Computer Science and Engineering Vishwakarma Institute of Technology, P une (INDIA) ratna.patil@vit.edu Dr. SHARVARI CHANDRASHEKHAR TAMANE Computer Science and Engineering Jawaharlal Nehru Engineering College, Aurangabad (INDIA) sharvaree73@yahoo.com Abstract—Diabetes Mellitus (DM) is a metabolic disease where the person will have high blood sugar due to the pancreas unable to produce sufficient insulin or the cells which are not responding to the insulin produced. There are three types of diabetes. They are Type 1 diabetes, Type 2 diabetes and Gestational diabetes. Type 1 diabetes is mostly occurring in children. Type 2 diabetes is called adult-onset diabetes which is common in adults. Gestational diabetes is only in women during pregnancy. Disease diagnosis is one of the applications where machine learning algorithms are giving successful results. This paper identifies gaps in the research on Type 2 diabetes disease diagnosis and treatment. Different classifiers can be used to explore patients ’ data and extract a predictive model. The importance of early diagnosis associated with the appropriate treatment is to decrease the chance of developing further complications like nerve damage, kidney failure, heart disease, diabetic retinopathy. Machine learning algorithms can provide reliable performance in determining diabetes mellitus. The focus of this paper is to study the recent algorithms used for diagnosis of Type 2 diabetes and find the research gaps. Keywords- Diabetes mellitus; early diagnosis; GA; machine learning; classification. I. INTRODUCTION Diabetes is a deadly disease and a major public health challenge worldwide. The number of diabetics in India is doubled from 32 million in 2000 to 63 million in 2013 and the figure is projected to further increase to 102.2 million in the next 15 years. This is the latest assessment by the World Health Organization, raising an alarm over the need to treat the condition. The annual spend on diabetes treatment in India is pegged at Rs1.5 lakh crore, which is 4.7 times the Centre’s allocation of Rs32000 crore for health. This cost is projected to rise by 20-30% every year. Diabetes disease diagnosis via proper interpretation of the Diabetes data is an important classification problem. There are several methodologies available on classification of diabetes disease. But less work has been done on early detection of the disease. This work will help to develop a predictive model based on set of attributes collected from the patients to develop a mathematical model. It is essential to find a way that can help in early detection with high accuracy and less complexity. MATERIAL A. DATASET In the machine learning research community, a work is going on to solve the classification problem. Pima Indian Dataset (PIMA) has been used to test the classification performance by most of the scholars. It is publicly available in the machine learning dataset UCI. All the instances in this dataset are Pima Indian women of at least 21 years old and living near Phoenix, Arizona, USA. The data is a collection of 768 records. B. Risk Factors The following are the parameters which contributes to the development of diabetes. The prevalence of Type 2 diabetes is increasing at a fast pace due to obesity, physical inactivity and unhealthy dietary habits.  Age – Indians develop diabetes earlier than western population. An early occurrence gives abundant time for the development of prolonged complications of diabetes. The incidence of diabetes increases with an age.  Family History – The occurrence of diabetes increases with a family history of diabetes. A high incidence of diabetes is seen among the first degree relatives.  Lifestyle – Deskbound lifestyle is an independent factor for the growth of Type2 diabetes.  Obesity – There is a close association of obesity with Type2 diabetes. Increase in weight increases Body Mass Index (BMI).  Stress – The impact of physical and mental stress along with lifestyle changes has an effect of incidence of Type 2 diabetes from persons in a strong genetic background International Journal of Computer Science and Information Security (IJCSIS), Vol. 14, No. 10, October 2016 329 https://sites.google.com/site/ijcsis/ ISSN 1947-5500