Blood Diseases Detection Using Data Mining Techniques

Al-Ameri M. A., CİYLAN B., Almassri S. D.

6th International Conference on Computer Science and Engineering, UBMK 2021, Ankara, Türkiye, 15 - 17 Eylül 2021, ss.73-77 identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.1109/ubmk52708.2021.9558942
  • Basıldığı Şehir: Ankara
  • Basıldığı Ülke: Türkiye
  • Sayfa Sayıları: ss.73-77
  • Anahtar Kelimeler: Component, Formatting, Insert Introduction, Style, Styling
  • Gazi Üniversitesi Adresli: Evet

Özet

© 2021 IEEEBy Using data mining tecliniques, mucli study lias been undertaken to derive information from medical data, which may enhance the curing ability. This study has been compared four automated methods for blood diseases detection using a different method for attributes reduction methods: PCA (Principal Component Analysis), ICA (Independent Component Analysis), original data set, and four algorithms, which are Naive Bayes, K-Nearest Neighbor, Decision Tree, and SVM (Support Vector Machine) for classification. The results show that after applying the PCA technique, the accuracy is: 0.98, 1.0, 0.79, and 0.78 for Naive Bayes, Decision Tree, SVM, and KNN and after applying the ICA technique, the accuracy is 0.76, .76, 0.76, and 0.74 for Naive Bayes, Decision Tree, SVM, and KNN, while the accuracy after applying on the original data set is: 0.98, 1.0, 0.78, and 0.75 for Naive Bayes, Decision Tree, SVM, and KNN. The results appeared that when applying decision tree on the original data set and after applying the PCA method, have been gotten the highest accuracy which is 100%.