| Title |
Prediction of Diabetes using IQR Algorithm of Interquartile Range Adjustment |
| Authors |
홍경찬(Kyeong-Chan Hong) ; 장명수(Myeong-su Jang) ; 김창민(Chang-Min Kim) ; 이우범(Woo-Beom Lee) ; 홍유식(Yoo-Sik Hong) ; 한영환(Young-Hwan Han) |
| DOI |
https://doi.org/10.5573/ieie.2023.60.3.31 |
| Keywords |
Diabetes; IQR algorithms; InterQuartile Range; Outliers; SMOTE; Imbalances class |
| Abstract |
In this paper, a method of pretreatment of diabetes data used in machine learning is proposed to diagnose and predict diabetes early. The preprocessing methods include missing values, abnormal values, and level imbalance, which are representative problems in medical data. In particular, the IQR algorithm, which is used to solve the ideal value problem, considers a large amount of data as ideal value when the data gap is large, so this paper proposes an ideal value elimination algorithm suitable for input data. The median and SMOTE algorithms are used to solve the problem of missing values and grade imbalance. The proposed pretreatment method shows 5% better performance than the previous ideal value algorithm, and the AUC value is also 4%. |