KNOWLEDGE DISCOVERY FROM DATA SETS THROUGH DATA MINING TECHNIQUES: APPLICATION TO MEDICAL DATA MINING

Year 2011, Volume: 6 Issue: 1, 237 - 245, 01.02.2011

Ömer Akgöbek Serkan Kaya

Abstract

A rapid development and growing use of database systems, and the importance of information stored in these systems raise the issue of how to make the best use of these systems. One of the leading area where the database systems are mostly used is 'medicine'. Today, all patients' laboratory results, patient history as well as x-ray images and more are kept in databases. These of the traditional database query and report methods to filter information and to present reports does not always provide the important hidden rules contained in the stored data. Therefore, the use of data mining techniques used for knowledge discovery in this area from databases is inevitable. In this study, the rules base is created thorough the knowledge discovery by employing REX-1 algorithm, a data mining technique, on the Wisconsin Breast Cancer, Ljubljana Breast Cancer, Dermatology, Hepatitis and Diabetes sample sets, which are real life data and commonly used in the medical field. In terms of the accuracy rate, the results of this study were compared to the results of the algorithms widely used in this field, such as C4.5, NavieBayes, PART, CN2, CORE, GA-SVM.

Keywords

Data mining, Knowledge-discovery, Medical data mining, Classification, .,

VERY MADENCYLY?Y TEKNYKLERY YLE VERY KÜMELERYNDEN BYLGY KE?FY: MEDYKAL VERY MADENCYLY?Y UYGULAMASI

Abstract

Veri tabany sistemlerinin büyük bir hyzla geli?mesi, artan kullanymy ve bu sistemlerdeki bilgilerin önemi bu sistemlerden nasyl yararlanylaca?y problemini de beraberinde getirmi?tir. Bu sistemlerin en çok kullanyldy?y alanlaryn ba?ynda da 'typ' gelmektedir. Günümüzde hastalara ait tüm laboratuar sonuçlary, hastanyn hikâyesi gibi bilgilerin yany syra çekilen film ve röntgen görüntüleri dahi veri tabanlarynda tutulmaktadyr. Bu veri tabanlaryndan geleneksel sorgulama metotlaryyla bilgiyi süzmek ve bu bilgileri raporlar halinde sunmak bilgiler içerisinde sakly bulunan gizli-önemli kurallaryn ortaya çykmasyny sa?lamaz. Bundan dolayy veri tabanlaryndan bilgi ke?fi için bu alanda kullanylan veri madencili?i tekniklerinin kullanylmasyny kaçynylmaz yapmaktadyr. Bu çaly?mada veri madencili?i tekniklerinden REX-1 algoritmasy kullanylarak medikal alanda kullanylan ve gerçek hayattan alynan Wisconsin Breast Cancer, Ljubljana Breast Cancer, Dermatology, Hepatitis ve Diabetes örnek setleri üzerinde bilgi ke?fi yapylarak kural tabany olu?turulmu?tur. Elde edilen sonuçlar bu alanda yaygyn olarak kullanylan C4.5, NavieBayes, PART, CN2, CORE, GA-SVM gibi algoritmalarla do?ruluk oranlaryna göre test edilmi?tir.

Keywords

Veri madencili?i, Bilgi ke?fi, Medikal veri madencili?i, Synyflandyrma, .,