• Login
    View Item 
    •   USU-IR Home
    • Faculty of Computer Science and Information Technology
    • Department of Information Technology
    • Master Theses
    • View Item
    •   USU-IR Home
    • Faculty of Computer Science and Information Technology
    • Department of Information Technology
    • Master Theses
    • View Item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    Analisis Kinerja K-Nearest Neighbor Menggunakan Principal Component Analysis (PCA)

    View/Open
    Fulltext (4.552Mb)
    Date
    2020
    Author
    Lubis, Aulia Hadi
    Advisor(s)
    Sihombing, Poltak
    Nababan, Erna Budhiarti
    Metadata
    Show full item record
    Abstract
    One of the weaknesses of the classification technique in the K-NN method is the level of accuracy that is still not optimal, this is supported by many features that are less relevant to the dataset so that it affects the level of classification changes, because it requires a method to reduce the features that are lacking. The relevant so that it is expected to improve the classification accuracy of the K-NN method. Performance that affects the weaknesses in the KNN classification method in the data grouping process often occurs. This process is intended to increase the results seen in proportion to accuracy. This study shows a comparison between the conventional KNN classification model with the PCA + KNN model against the Air Quality dataset where the accuracy value generated from the model PCA + KNN is always higher than the conventional K-NN model, which is 93.06% compared to the The accuracy of Conventional K-NN is 90.74% when the K value is simultaneously at the K = 5 value. While the lowest value can be obtained in the Conventional K-NN model, which is 87.96%, when the K value is at the K = value. 4. Thus, the accuracy value of the PCA + K-NN model has a significant advantage of 2.32% over the Conventional K-NN model for the Air Quality dataset.
     
    Salah satu kelemahan dari teknik klasifikasi pada metode K-NN adalah tingkat keakurasian nya yang masih belum optimal, hal ini dipengaruhi oleh banyak nya fitur yang kurang relevan pada dataset sehingga mempengaruhi tingkat akurasi peng-klasifikasi-annya, oleh karena itu diperlukan metode untuk mengurangi fitur-fitur yang kurang relevan sehingga diharapkan mampu meningkatkan akurasi pengklasifikasian metode K-NN. Performance yang berpengeruh terhadap Kelemahan pada metode klasifikasi KNN pada proses pengelompokkan data sering terjadi. Proses ini dimaksud dalam peningkatan pencapaian hasil yang terlihat dalam persentase akurasi. Penelitian ini menunjukkan perbandingan akurasi antara model klasifikasi.KNN.Konvensional dengan model.PCA+KNN terhadap dataset Kualitas Udara dimana nilai.akurasi yang.dihasilkan dari model.PCA+KNN selalu lebih tinggi daripada model K-NN Konvensional, yakni sebesar 93.06% dibanding dengan akurasi K-NN Konvensional sebesar 90.74% saat.nilai K serentak berada.pada.nilai K=5. Sementara nilai akurasi terendah diperoleh pada model K-NN Konvensional yakni sebesar 87.96%, saat nilai K berada pada nilai K=4. Dengan demikian, nilai akurasi pada model PCA+K-NN memiliki keunggulan akurasi yang signifikan 2.32% terhadap model K-NN Konvensional untuk dataset Kualitas Udara.

    URI
    http://repositori.usu.ac.id/handle/123456789/30375
    Collections
    • Master Theses [621]

    Repositori Institusi Universitas Sumatera Utara - 2025

    Universitas Sumatera Utara

    Perpustakaan

    Resource Guide

    Katalog Perpustakaan

    Journal Elektronik Berlangganan

    Buku Elektronik Berlangganan

    DSpace software copyright © 2002-2016  DuraSpace
    Contact Us | Send Feedback
    Theme by 
    Atmire NV
     

     

    Browse

    All of USU-IRCommunities & CollectionsBy Issue DateTitlesAuthorsAdvisorsKeywordsTypesBy Submit DateThis CollectionBy Issue DateTitlesAuthorsAdvisorsKeywordsTypesBy Submit Date

    My Account

    LoginRegister

    Repositori Institusi Universitas Sumatera Utara - 2025

    Universitas Sumatera Utara

    Perpustakaan

    Resource Guide

    Katalog Perpustakaan

    Journal Elektronik Berlangganan

    Buku Elektronik Berlangganan

    DSpace software copyright © 2002-2016  DuraSpace
    Contact Us | Send Feedback
    Theme by 
    Atmire NV