Analisis Potensi Pendonor Darah di Unit Transfusi Darah Palang Merah Indonesia (UTD PMI) Kota Surabaya dengan Metode Classification Tree dan Neural Network

Salam, Fakhrus (2017) Analisis Potensi Pendonor Darah di Unit Transfusi Darah Palang Merah Indonesia (UTD PMI) Kota Surabaya dengan Metode Classification Tree dan Neural Network. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of ACCESS CLOSED BY AUTHOR] Text (ACCESS CLOSED BY AUTHOR)
1315105026-Undergraduate_Theses.pdf - Published Version
Restricted to Repository staff only

Download (2MB) | Request a copy

Abstract

Persediaan darah dapat berubah sewaktu-waktu, sementara jumlah permintaan darah dari pasien terus meningkat. Pada saat kekurangan persediaan darah, UTD PMI Kota Surabaya menghubungi beberapa orang secara acak untuk melakukan donor darah, metode tersebut dirasa masih kurang efektif karena data pendonor yang terkumpul seringkali tidak memenuhi syarat donor darah. Analisis potensi pendonor darah dapat dijadikan sebagai salah satu tindakan pengambilan keputusan. Salah satu metode untuk pengambilan keputusan sistematis adalah pohon keputusan seperti classification tree dengan kelebihan dapat digunakan untuk data kategorik dan numerik serta menangani adanya data yang tidak lengkap. Selain itu, metode neural network (NN) dipilih karena memiliki kelebihan tanpa adanya asumsi dan mampu melakukan generalisasi dari data. Data yang digunakan yaitu data transaksi donor darah di UTD PMI Kota Surabaya pada Januari - Desember 2016, dengan variabel respon yaitu donor lagi (ya dan tidak) serta variabel prediktor yaitu jenis kelamin, usia, tekanan darah sistolik dan diastolik, kadar HB dan wilayah pendonor. Performa terbaik dari hasil klasifikasi pendonor darah adalah metode NN. Accuracy data testing untuk golongan darah A yaitu 55,76 persen, golongan darah AB sebesar 55,40, golongan darah B sebesar 55,59 persen, dan golongan darah O sebesar 55,82 persen.
=============================================================================Blood supply can change at any time while the amount of blood demand from patients continues to increase. If UTD PMI Surabaya has lack of blood’s supply, they contacted people randomly to donor. Aforementioned method is considered to be less effective because the people often do not meet the requirements to blood donor. Analysis of potential blood donors can be used as one of decision making. One of the methods for systematic decision-making is decision tree such as classification tree which can be used for categorical and numerical data and able to accommodate incomplete data. In addition, the neural network method was chosen because of its no assumptions and ability to generalization of data. Data used for this analysis are blood donor transaction data of UTD PMI Surabaya in January - December 2016, with the will to donor more (yes and no) as the response variable and several predictor variables such as gender, age, systolic and diastolic blood pressure, HB, and donor’s area. The best performance of the blood donors classification is the neural network method. The accuracy level for A type blood’s testing was 55.76 percent, 55.40 percent for AB type blood, 55.59 percent for B type blood, and 55.82 percent for O type blood.

Item Type: Thesis (Undergraduate)
Additional Information: RSSt 519.536 Sal a
Uncontrolled Keywords: classification tree, neural network, pendonor darah, UTD PMI Kota Surabaya
Subjects: H Social Sciences > HA Statistics
Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions: Faculty of Mathematics and Science > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User: Fakhrus Salam .
Date Deposited: 13 Feb 2018 07:52
Last Modified: 11 Sep 2023 07:11
URI: http://repository.its.ac.id/id/eprint/47803

Actions (login required)

View Item View Item