Open AccessSISTEM PENDUKUNG KEPUTUSAN UNTUK DETEKSI DINI RISIKO PENYAKIT STROKE MENGGUNAKAN LEARNING VECTOR QUANTIZATION
Author(s) -
Sugarwanto Atmaja
Publication year - 2019
Publication title -
indexia
Language(s) - English
Resource type - Journals
eISSN - 2657-0432
pISSN - 2657-0424
DOI - 10.30587/indexia.v1i1.823
Subject(s) - learning vector quantization , blood pressure , medicine , blood sugar , uric acid , stroke (engine) , affect (linguistics) , creatinine , artificial neural network , physical therapy , machine learning , computer science , psychology , endocrinology , diabetes mellitus , mechanical engineering , communication , engineering
Stroke is any sudden neurologic disorder that occurs as a result of restriction or cessation of blood flow through brain arteries supply system. Stroke is the leading cause of death in Indonesia. Services pre-stroke is early detection activities, discovery and monitoring of risk factors for stroke in healthy individuals and at-risk communities that can be performed by physicians, nurses and health workers. Based on the results of the study said that when control stroke risk factors to do with the approach would reduce the number of defects by 60-90%. Works doctor for diagnosis process is not easy because of the many risk factors vary and affect each other, for example Low Density Lipoprotein cholesterol can lead to heart disease can also affect blood pressure, gender can affect the value of uric acid, uric acid can also influence blood pressure and sugar levels can affect blood pressure. Classification method is one solution that is deemed able to handle the process of classifying the status of early detection of the risk of stroke. Mechanical classification using Learning Vector Quantization (LVQ) has excess generating an error value is smaller than other artificial neural networks. Based on the results of research and discussion conducted, algorithms LVQ can recognize patterns and are able to predict the status of early risk of stroke using a variable blood pressure, blood sugar, total cholesterol, Low Density Lipoprotein, age, gender, gout, Blood Urea Nitrogen and creatinine with a total value of up to 82% accuracy.