Introduction PROS Ria DLNK, Alexandr K, Heri K Random forest fulltext
Proceedings of the IConSSE FSM SWCU 2015, pp. MA.26–41 ISBN: 978-602-1047-21-7
SWUP
MA.26
Random forest of modified risk factor on ischemic and hemorrhagic
Case study: Medicum Clinic, Tallinn, Estonia
Ria Dhea Layla Nur Karisma
a
, Alexandr Kormitsõn
b
, Heri Kuswanto
c
a,c
Sepuluh Nopember Institute of Technology, Jl. Arief Rahman Hakim, Surabaya 60117, Indonesia
b
Tallinn University of Technology, Ehitaja Tee 5, Tallinn 19086 , Estonia
Abstract
Estonia is one of European Union countries with capital city named Tallinn. It is one of Baltic area with population 1312300 and they have problem in health such as Stroke
Cerebrovascular which is the second biggest cardiovascular disease cause of death. The aim of study is to classify modified factor Ischemic patient and Hemorrhagic Patient
using ensemble method. It used Random Forest analysis which is a classifier formed from a set of tree structure, where each tree is a random independent vector which has
identical distribution and each tree comes from best unit. Generally, the method has better accuracy than individual classification. The unit of observation is 420 patients
consist of missing data and the independent variable is modified factor of Ischemic patient and Hemorrhagic patient in Medicum clinic, Tallinn, Estonia. The independent
variable is alcohol habit, diet habit, smoking habit, physical activity, and body mass index. Proportion of training and testing data is 85:15, whereas it formed proportion
of original data set. In this research, used bootstrap with replacement 2015 times one used and replication 300 along 3 combination of predictor variable, which is 1,7 in miss
accuracy. The important modified risk factor is diet habit and alcohol habit. Variable that has influenced in Ischemic is smoking habit, diet habit, and physical activity meanwhile
in Hemorrhagic is diet habit. Response variable has imbalance data then we are considered for appropriate accuracy that showed by sensitivity and specificity. Accuracy
of prediction model 98.32 and validation of the model is 95.23, then sensitivity and specificity are 98.6 and 97.2 respectively.
Keywords stroke, ischemic, hemorrhagic, modified risk factor, ensemble method, random forest