KOMBINASI METODE IMPUTASI MEAN DAN MULTIPLE IMPUTATION BY CHAINED EQUATIONS (MICE) UNTUK PENANGANAN DATA HILANG DAN PENINGKATAN EVALUASI KINERJA KLASIFIKASI PREDIKSI PENYAKIT DIABETES MELITUS

TASYA, YULFITA and Desiani, Anita and Andriani, Yuli (2023) KOMBINASI METODE IMPUTASI MEAN DAN MULTIPLE IMPUTATION BY CHAINED EQUATIONS (MICE) UNTUK PENANGANAN DATA HILANG DAN PENINGKATAN EVALUASI KINERJA KLASIFIKASI PREDIKSI PENYAKIT DIABETES MELITUS. Undergraduate thesis, Sriwijaya University.

	Text RAMA_44201_08011281924041.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (1MB) | Request a copy
	Text RAMA_44201_08011281924041_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (7MB) | Request a copy
Preview	Text RAMA_44201_08011281924041_0011127702_0002077202_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (807kB) | Preview
	Text RAMA_44201_08011281924041_0011127702_0002077202_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (432kB) | Request a copy
	Text RAMA_44201_08011281924041_0011127702_0002077202_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (360kB) | Request a copy
	Text RAMA_44201_08011281924041_0011127702_0002077202_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (552kB) | Request a copy
	Text RAMA_44201_08011281924041_0011127702_0002077202_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (85kB) | Request a copy
	Text RAMA_44201_08011281924041_0011127702_0002077202_06_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (213kB) | Request a copy

Abstract

Pima Indians Diabetes 2020 dataset is one of the datasets that contains missing data. Missing data can cause some statistical information to be lost due to the small sample size and can cause overfitting problems in the training data. One way to deal with missing data can be done by imputing data. This study aims to improve classification performance on Pima Indians Diabetes 2020 dataset by applying a combination of Single Imputation using the Mean imputation method on attributes containing missing data less than or equal to 10% and Multiple Imputation using MICE on attributes containing more than 10% missing data. 10%. The results of missing data imputation were tested using the Multi Layer Perceptron (MLP) and Support Vector Machine (SVM) methods to find out the increase in classification performance evaluation. Before handling missing data, the results of the classification performance evaluation obtained an accuracy of 78.947%, a precision of 78.554%, and a recall of 76.616%, after handling missing data using the Mean and MICE methods, the results of the classification performance evaluation obtained an accuracy of 84.221%, a precision of 82.462%, and a recall of 82.462%. Accuracy, precision and recall values increased by 5.274%, 3.908% and 5.846% respectively. It can be concluded that the prediction of missing data using the Multi Layer Perceptron (MLP) and Support Vector Machine (SVM) methods can improve the performance evaluation of the prediction classification of diabetes mellitus.

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	Data Hilang, Mean, MICE, MLP, SVM, dan Klasifikasi
Subjects:	Q Science > Q Science (General) > Q334-342 Computer science. Artificial intelligence. Algorithms. Robotics. Automation.
Divisions:	08-Faculty of Mathematics and Natural Science > 44201-Mathematics (S1)
Depositing User:	Yulfita Tasya
Date Deposited:	22 Feb 2023 08:27
Last Modified:	22 Feb 2023 08:27
URI:	http://repository.unsri.ac.id/id/eprint/89866

Actions (login required)

View Item