Local cover image
Local cover image
Local cover image
Local cover image

D-neighborhood imputation method for ordinal data sets with missing values / Jon Marx P. Sarmiento

By: Material type: TextTextLanguage: English Publication details: 2007Description: 111 leavesSubject(s): Abstract: Imputation is applied in filling up missing values in surveys which are ordinal in form. Among the imputation techniques are Mean, Mode, Hot-deck and KNN imputations which have their own drawbacks. To address this issue, the proponent introduced a new imputation method called D-neighborhood imputation. It uses the concept of neighborhood and cut off value to ensure high similarity with the reference and the maximum penalty rule in solving for the distance of unknown values. D-neighborhood was evaluated and compared with the existing techniques. The experiment was done using the Dermatology and Breast Cancer data sets. Incomplete data sets were generated under MCAR with 1%, 5%, 10%, 20%, and 30% level of missing values and conditioned MCAR with 0.25, 0.5, 0.75 and 1 probability in no, 2, and 3 combinations. According to the results, it performed best under MCAR condition in both data sets and resulted the best clustering quality when applied to Breast Cancer data set under MAR condition. Using Dermatology data set, D-neighborhood and KNN have competing results while using Breast Cancer data set, D-neighborhood performed best. In general, D-neighborhood imputation outperformed the rest of the algorithms when tested in both data sets.
List(s) this item appears in: BS Applied Mathematics
Tags from this library: No tags from this library for this title. Log in to add tags.
Star ratings
    Average rating: 0.0 (0 votes)
Holdings
Cover image Item type Current library Collection Call number Status Date due Barcode
University Library Theses Room-Use Only LG993.5 2007 A64 S27 (Browse shelf(Opens below)) Not For Loan 3UPML00012083
University Library Archives and Records Preservation Copy LG993.5 2007 A64 S27 (Browse shelf(Opens below)) Not For Loan 3UPML00035015

Thesis (BS Applied Mathematics) -- University of the Philippines Mindanao, 2007

Imputation is applied in filling up missing values in surveys which are ordinal in form. Among the imputation techniques are Mean, Mode, Hot-deck and KNN imputations which have their own drawbacks. To address this issue, the proponent introduced a new imputation method called D-neighborhood imputation. It uses the concept of neighborhood and cut off value to ensure high similarity with the reference and the maximum penalty rule in solving for the distance of unknown values. D-neighborhood was evaluated and compared with the existing techniques. The experiment was done using the Dermatology and Breast Cancer data sets. Incomplete data sets were generated under MCAR with 1%, 5%, 10%, 20%, and 30% level of missing values and conditioned MCAR with 0.25, 0.5, 0.75 and 1 probability in no, 2, and 3 combinations. According to the results, it performed best under MCAR condition in both data sets and resulted the best clustering quality when applied to Breast Cancer data set under MAR condition. Using Dermatology data set, D-neighborhood and KNN have competing results while using Breast Cancer data set, D-neighborhood performed best. In general, D-neighborhood imputation outperformed the rest of the algorithms when tested in both data sets.

There are no comments on this title.

to post a comment.

Click on an image to view it in the image viewer

Local cover image Local cover image
 
University of the Philippines Mindanao
The University Library, UP Mindanao, Mintal, Tugbok District, Davao City, Philippines
Email: library.upmindanao@up.edu.ph
Contact: (082)295-7025
Copyright @ 2022 | All Rights Reserved