Juwita, Ratna (2018) Klasifikasi Kelas Risiko Pasien Pneumonia Menggunakan Regresi Logistik Ordinal, Hybrid Regresi Logistik Ordinal – Algoritma Genetika dan Naïve Bayes Classification. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
1314100091-Undergraduate_Theses.pdf - Accepted Version Download (2MB) | Preview |
Abstract
Pneumonia merupakan salah satu infeksi saluran napas akut yang menjadi penyebab tingginya angka kematian di Indonesia. Pada umumnya tim medis akan melakukan diagnosis awal terhadap pasien kemudian menentukan derajat keparahan pneumonia yang diderita pasien sebagai upaya untuk mengetahui penanganan yang tepat bagi pasien. Penilaian ini menggunakan sistem skor Pneumonia Severty Index (PSI) dan sistem skor CURB-65, namun belum ada acuan yang pasti untuk menentukan sistem penilaian yang akan digunakan. Oleh karena itu perlu dilakukan seleksi variabel untuk menentukan variabel-variabel yang berpe-ngaruh dalam penentuan kelas risiko pneumonia. Pada penelitian ini digunakan seleksi variabel menggunakan metode backward, forward dan stepwise. Selain itu digunakan pendekatan algoritma genetika untuk seleksi variabel dan optimasi parameter pada Regresi Logistik Ordinal dan Naive Bayes Classification untuk mengelompokkan kelas risiko pasien pneumonia di RSUD Dr. Soetomo. Hasil klasifikasi menunjukkan bahwa metode Hybrid Regresi Logistik Ordinal-Algoritma Genetika adalah metode yang paling baik karena menghasilkan nilai ketepatan klasifikasi tertinggi.
===================================================================================================================
Pneuomonia is one of the acute respiratory infections that cause the high mortality rate in Indonesia. In general, the medical team will make an initial diagnosis of the patient then the medical team will assess the degree of severity of pneumonia suffered by the patient in an effort to determine the appropriate treatment for the patient. This assessment uses the Pneumonia Severty Index (PSI) and the CURB-65 score system but there is no definitive reference for determining the system assessment to be used. Therefore, variable selection is needed to determine the variables that influence the determination of pneumonia risk class. This research applies Ordinal Logistic Regression method with selection of forward selection, backward elimination and stepwise method. In addition, genetic algorithm approach is used for variable selection and parameter optimization for Ordinal Logistic Regression and Nai ̈ve Bayes Classification to classify risk class of pneumonia patient in RSUD Dr. Soetomo. The classification results show that Hybrid Ordinal Logistic Regression - Genetic Algorithm is the best method because it produces the highest accuracy.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Additional Information: | RSSt 519.536 Juw k-1 3100018074475 |
| Uncontrolled Keywords: | Algoritma Genetika; Ketepatan Klasifikasi; Naïve Bayes; Pneumonia; Regresi Logistik Ordinal; Accuracy; Genetic Algorithm; Naïve Bayes; Pneumonia; Ordinal Logistic Regression |
| Subjects: | H Social Sciences > HA Statistics Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression Q Science > QA Mathematics > QA402.5 Genetic algorithms. Interior-point methods. R Medicine > RB Pathology |
| Divisions: | Faculty of Mathematics and Science > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Ratna Juwita |
| Date Deposited: | 09 Apr 2018 03:25 |
| Last Modified: | 21 Sep 2020 05:49 |
| URI: | http://repository.its.ac.id/id/eprint/50672 |
Actions (login required)
 |
View Item |