Research Title |
Missing Value Imputation based on K Nearest Neighbor Method with Correlation Coefficient |
Date of Distribution |
18 November 2018 |
Conference |
Title of the Conference |
INTERNATIONAL CONFERENCE ON APPLIED STATISTICS (ICAS) 2018 |
Organiser |
คณะวิทยาศาสตร์และเทคโนโลยี มหาวิทยาลัยสงขลานครินทร์ วิทยาเขตปัตตานี |
Conference Place |
Centra by Centara, Government Complex Hotel & Convention Centre Chaeng Watthana Bangkok |
Province/State |
กรุงเทพมหานคร |
Conference Date |
24 October 2018 |
To |
26 October 2018 |
Proceeding Paper |
Volume |
- |
Issue |
- |
Page |
54-57 |
Editors/edition/publisher |
|
Abstract |
This study is to propose a method for issing data imputation, namely Correlated K Nearest Neighbor (Corr-KNN) which prepare the data for analysis.
The correlation coefficient is used to select the variables having complete data which highly correlate with the variable having missing data. After
that, the selected data set in the variables having complete data and the missing data are used in K Nearest Neighbor (KNN) method by replacing
missing data with substituted data. The Corr-KNN method is compared with the KNN and K Nearest Neighbor Feature Selection (KNNFS) methods
by using the data on UCI Machine Learning Repository database. The performance of each method is measured by the Root Mean Squared Error
(RMSE). The results indicate that the Corr-KNN method has powerful imputation with smaller RMSE than the compared methods. |
Author |
|
Peer Review Status |
มีผู้ประเมินอิสระ |
Level of Conference |
นานาชาติ |
Type of Proceeding |
Full paper |
Type of Presentation |
Oral |
Part of thesis |
true |
Presentation awarding |
false |
Attach file |
|
Citation |
0
|
|