Rindrasari, Rahajeng (2018) Klasifikasi Kelas Risiko Pasien Pneumonia Menggunakan Metode Support Vector Machine-Genetic algotihm(SVM-GA) Hybrid. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
1314100117_Undergraduate Theses.pdf - Accepted Version Download (2MB) | Preview |
Abstract
Pneumonia, atau sering disebut radang paru-paru, dapat menimbulkan kerusakan pada jaringan paru-paru dan gangguan pertukaran gas setempat sehingga pneumonia dapat mengakibatkan sakit yang parah bahkan sampai berujung kematian. Dibutuhkan pengklasifikasian penyakit pneumonia untuk mempercepat dalam menentukan tingkat keparahan penyakit serta mengetahui tindakan yang paling tepat untuk penderita. Support Vector Machine dapat mengatasi permasalahan klasifikasi yang linier maupun tidak linier. Dalam penelitian ini, metode Genetic Algorithm dipakai dalam optimasi parameter. Penelitian ini menerapkan metode Grid Search Support Vector Machine dengan seleksi variabel dengan FCBF dan GA, serta Support Vector Machine-Genetic Algorithm (SVM-GA) Hybrid untuk mengklasifikasi kelompok kelas risiko berdasarkan rekam data medis pasien kemudian metode-metode tersebut dibandingkan nilai akurasinya. Hasil analisis menunjukkan bahwa metode SVM-GA dengan seleksi variabel GA memberikan nilai ketepatan klasifikasi yang lebih tinggi dibandingkan metode SVM dengan tanpa seleksi variabel serta seleksi variabel GA. Selain itu, optimasi parameter menggunakan metode GA dapat meningkatkan nilai akurasi pada data. ============= Pneumonia, or it is often called pneumonia, can cause damage to lung tissue and local gas exchange disturbances so that pneumonia can cause severe pain even to death. The classification of pneumonia disease is needed to accelerate in determining the severity of the disease as well as knowing the most appropriate action for the patient. Support Vector Machine can solve linear or non-linear classification problem. In this research, Genetic Algorithm method is used in parameter optimization. In this research applies Grid Search Support Vector Machine method with variable selection with FCBF and GA, and Support Vector Machine-Genetic Algorithm Hybrid (SVM-GA) to classify risk class group based on patient medical data record then those methods are compared the value of accuracy. The results of the analysis show that the SVM-GA method with GA variable selection gives a higher classification precision value than the SVM method with no variable selection and GA variable selection. In addition, parameter optimization using the GA method can improve the accuracy value of the data.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Additional Information: | RSSt 519.535 Rin k |
| Uncontrolled Keywords: | Genetic algorithm; optimasi; pneumonia; seleksi variabel; support vector machine |
| Subjects: | Q Science Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) Q Science > QA Mathematics > QA76.9 Computer algorithms. Virtual Reality. Computer simulation. |
| Divisions: | Faculty of Mathematics, Computation, and Data Science > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Rahajeng Rindrasari |
| Date Deposited: | 26 Jan 2018 03:47 |
| Last Modified: | 23 Sep 2020 03:43 |
| URI: | http://repository.its.ac.id/id/eprint/50653 |
Actions (login required)
 |
View Item |