Please use this identifier to cite or link to this item: http://repository.ipb.ac.id/handle/123456789/166948
Full metadata record
DC FieldValueLanguage
dc.contributor.advisorRuhiyat-
dc.contributor.advisorAnisa, Rahma-
dc.contributor.authorMahiradewi, Nayla Jasmine-
dc.date.accessioned2025-08-07T06:34:27Z-
dc.date.available2025-08-07T06:34:27Z-
dc.date.issued2025-
dc.identifier.urihttp://repository.ipb.ac.id/handle/123456789/166948-
dc.description.abstractStroke merupakan kondisi medis serius yang dapat menyebabkan kecacatan permanen atau bahkan kematian jika tidak ditangani dengan cepat. Penelitian ini menggunakan data sekunder dari dataset stroke yang terdiri dari 5,110 observasi, mencakup informasi mengenai usia, jenis kelamin, riwayat hipertensi, penyakit jantung, kadar glukosa rata-rata, indeks massa tubuh, tipe tempat tinggal, tipe pekerjaan, dan riwayat merokok. Dua metode machine learning, random forest dan XGBoost, dipilih untuk memprediksi risiko stroke. Kedua metode tersebut dianggap cocok untuk data besar dan lebih fleksibel karena tidak memerlukan asumsi parametrik. Penelitian ini bertujuan untuk membandingkan kinerja kedua metode tersebut dalam memprediksi risiko stroke. Evaluasi model dilakukan menggunakan metrik balanced accuracy, sensitivitas, spesifisitas, dan F1-score. Hasil penelitian menunjukkan bahwa XGBoost unggul pada balanced accuracy dibandingkan random forest. Analisis variable importance pada model XGBoost mengidentifikasi usia, rata-rata kadar glukosa, dan indeks massa tubuh sebagai faktor paling berpengaruh terhadap risiko stroke. Penelitian ini diharapkan dapat membantu pengembangan sistem pendukung keputusan untuk deteksi dini stroke serta memberikan wawasan mengenai faktor risiko utama yang perlu diperhatikan.-
dc.description.abstractStroke is a serious medical condition that can lead to permanent disability or even death if not promptly managed. This study utilizes secondary data from a stroke dataset comprising 5,110 observations, including information on age, gender, history of hypertension, heart disease, average glucose level, body mass index, residence type, work type, and smoking history. Two machine learning methods, random forest and XGBoost, were chosen to predict stroke risk. Both methods are considered suitable for large datasets and are more flexible as they do not require parametric assumptions. This research aims to compare the performance of these two methods in predicting stroke risk. Model evaluation was conducted using balanced accuracy, sensitivity, specificity, and F1-score metrics. The results indicated that XGBoost performs superiorly in terms of balanced accuracy compared to random forest. Variable importance analysis from the XGBoost model identified age, average glucose level, and body mass index as the most influential factors contributing to stroke risk. This study is expected to contribute to the development of decision support systems for early stroke detection and provide insights into key risk factors that need attention.-
dc.description.sponsorshipnull-
dc.language.isoid-
dc.publisherIPB Universityid
dc.titlePerbandingan Metode Machine Learning Random Forest dan XGBoost untuk Memprediksi Risiko Kejadian Strokeid
dc.title.alternativeThe Comparison of Random Forest and XGBoost Machine Learning Methods for Predicting Stroke Risk-
dc.typeSkripsi-
dc.subject.keywordrandom forestid
dc.subject.keywordMachine Learningid
dc.subject.keywordXGBoostid
dc.subject.keywordStrokeid
Appears in Collections:UT - Actuaria

Files in This Item:
File Description SizeFormat 
cover_G5402211067_a66d6dce86764169a4a1610ee31c2d54.pdfCover598.81 kBAdobe PDFView/Open
fulltext_G5402211067_c79776b2c243415d8ed921b1eb6b693d.pdf
  Restricted Access
Fulltext1.57 MBAdobe PDFView/Open
lampiran_G5402211067_670939b2b8184df5a08535b0b3f394b4.pdf
  Restricted Access
Lampiran191.23 kBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.