Sciweavers

Share
DAWAK
2009
Springer

Dynamic Clustering-Based Estimation of Missing Values in Mixed Type Data

11 years 8 months ago
Dynamic Clustering-Based Estimation of Missing Values in Mixed Type Data
The appropriate choice of a method for imputation of missing data becomes especially important when the fraction of missing values is large and the data are of mixed type. The proposed dynamic clustering imputation (DCI) algorithm relies on similarity information from shared neighbors, where mixed type variables are considered together. When evaluated on a public social science dataset of 46,043 mixed type instances with up to 33% missing values, DCI resulted in more than 20% improved imputation accuracy over Multiple Imputation, Predictive Mean Matching, Linear and Multilevel Regression, and Mean Mode Replacement methods. Data imputed by 6 methods were used for prediction tests by NB-Tree, Random Subset Selection and Neural Networkbased classification models. In our experiments classification accuracy obtained using DCI-preprocessed data was much better than when relying on alternative imputation methods for data preprocessing.
Vadim V. Ayuyev, Joseph Jupin, Philip W. Harris, Z
Added 19 May 2010
Updated 19 May 2010
Type Conference
Year 2009
Where DAWAK
Authors Vadim V. Ayuyev, Joseph Jupin, Philip W. Harris, Zoran Obradovic
Comments (0)
books