Data Collection, Statistical Analysis and Clustering Studies of Cancer Dataset from Viziayanagaram District, AP, India

Abstract

Cancer detection is one of major research that can be processed through datasets and data mining techniques. The data has been collected from Vizianagaram district (Village) during 2013 with 328 instances and 28 attributes (Gender, Age, Cancer Type, Family_members, Drinking, Smoking, Tea, Coffee, perfumes, Morning_eat, Travelling, Wake_up, Sleep, Tensions, Cool_drinks, Icecream, Height, weight, hair_loss, Marital, milk, bath, Oil, Fast_food, other diseases, Mobile, Sports, Mosquito_replents). The dataset has been analyzed using weka version 3.6.3and Orange softwares v2.7. The histogram shows higher instances for Lung cancer (56), Mouth (40), Bone (40), Skin (32), and Colon (24). There are more number of instances observed in Males (53.7%) compared with females (46.3%). The disease in married people are more (61%) compared to unmarried (39%) with average age groups observed at 33.78±10.12, Height as 159.02±9.79 cms and weight as 61.55±11.69 Kgs. Nearly 90.2% patients has no other diseases, 136 patients (41.5%) prefer drinking alcohol, 72 patients (22%) prefer smoking, 208(63.4%) prefer drinking tea, 96 (29.3%) prefer drinking coffee, 216(65%) prefer taking rice, 80(24.4%) prefer taking cool drinks, no person like ice creams, 88 (26.8%) prefer taking milk, 238(63.4%) prefer taking sunflower oil in cooking and 68(26.8%) prefer taking fast food. The data shows hair loss, use of mobile phones and mosquito repellents as major factors in cancer. It concludes that Age, Gender, Height, weight, marital status, tea, walking, hairloss, mobile and mosquito repellents are major factors/attributes in cancer occurrence. © Springer International Publishing Switzerland 2014.