Please use this identifier to cite or link to this item:
http://repository.ipb.ac.id/handle/123456789/158488
Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Ardana, Ngakan Komang Kutha | |
dc.contributor.advisor | Sumarno, Hadi | |
dc.contributor.author | Maulia, Syammira Dhifa | |
dc.date.accessioned | 2024-08-24T15:17:07Z | |
dc.date.available | 2024-08-24T15:17:07Z | |
dc.date.issued | 2024 | |
dc.identifier.uri | http://repository.ipb.ac.id/handle/123456789/158488 | |
dc.description.abstract | Data kompleks seringkali mengandung pencilan signifikan, yang terbagi menjadi pencilan baris dan pencilan sel. Penelitian ini berfokus pada pencilan sel, yang dapat menyebabkan kesalahan analisis jika tidak ditangani dengan baik. Analisis Komponen Utama (AKU) klasik sering terpengaruh oleh pencilan, sehingga diperlukan metode pengembangan seperti AKU Macro dan AKU Kekar. AKU Macro mampu menangani pencilan sel, pencilan baris, dan data yang hilang, sementara AKU Kekar menggabungkan Projection Pursuit dengan Minimum Covariance Determinant. Metode Detect Deviating Cells (DDC) secara khusus dapat mendeteksi dan menangani pencilan sel. Berdasarkan ukuran kesesuaian dalam menganalisis komponen utama, AKU Macro menunjukkan hasil yang lebih efektif dibandingkan metode lain. Selain itu, analisis dengan plot tebaran score distance dan orthogonal distance menunjukkan bahwa AKU Macro lebih efektif dalam mendeteksi pencilan. Hasil penelitian ini menegaskan bahwa AKU Macro meningkatkan keandalan analisis data yang kompleks, menawarkan pendekatan yang lebih robust dalam mengelola pencilan. | |
dc.description.abstract | Complex data often contains significant outliers, which can be categorized into rowwise and cellwise outliers. This study focuses on cellwise outliers, which can cause severe analytical errors if not properly addressed. Classical Principal Component Analysis (PCA) is highly susceptible to outliers, necessitating advanced methods such as Macro PCA and Robust PCA. Macro PCA can handle cellwise outliers, rowwise outliers, and missing data, while Robust PCA integrates Projection Pursuit with the Minimum Covariance Determinant. The Detect Deviating Cell (DDC) method is specifically designed to detect and manage cellwise outliers. Based on fit measures in principal component analysis, Macro-PCA is superior to other methods. Furthermore, analysis using score distance and orthogonal distance scatter plots demonstrates that Macro PCA is more effective in outlier detection. This research confirms that Macro PCA significantly enhances the reliability of complex data analysis, providing a robust approach to outlier management. | |
dc.description.sponsorship | ||
dc.language.iso | id | |
dc.publisher | IPB University | id |
dc.title | Efektivitas Reduksi Dimensi Data pada Analisis Komponen Utama dengan Penanganan Pencilan Sel | id |
dc.title.alternative | ||
dc.type | Skripsi | |
dc.subject.keyword | pencilan | id |
dc.subject.keyword | AKU Kekar | id |
dc.subject.keyword | AKU Klasik | id |
dc.subject.keyword | AKU Macro | id |
dc.subject.keyword | DDC | id |
Appears in Collections: | UT - Mathematics |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
cover_G5401201059_1e7b9bf385bd479fa0af5c33d31d46f8.pdf | Cover | 323.75 kB | Adobe PDF | View/Open |
fulltext_G5401201059_562daeb437274ec8aab5ef2b05113683.pdf Restricted Access | Fulltext | 2.48 MB | Adobe PDF | View/Open |
lampiran_G5401201059_88a38ff47bd54ac5a9a39054cda52899.pdf Restricted Access | Lampiran | 1.97 MB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.