A Comparative Study of Missing Value Imputation Methods on Time Series Data

Author(s):

Arumuga Nainar S.

Published in:

HCTL Open International Journal of Technology Innovations and Research (IJTIR), eISSN: 2321-1814

Published on:

02-May-2015

Volume:

Volume 14, April 2015, ISBN:978-1-62951-946-3.

Copyright Information:

© 2015 by the Authors; Licensed by HCTL Open, India.

License Information:

This article is an open-access article distributed under the terms and conditions of the Creative Commons Attribution 4.0 International License.

Abstract

Data mining is process of analyzing enormous sets of data and extracting meaningful information from data. Temporal data mining deals with data which has time information.In practical, the data collected may contain noisy, inconsistent data and in many cases the data may be missing. So one of the important step that need to be done in data mining is data pre-processing. Incomplete data may generate biased results and impact the accuracy of analysis. In order to rectify this it is important to predict the missing values based on other details in the dataset. The work focuses on predicting missing data using mean imputation, hotdecking and Inverse Distance Weighted Interpolation methods and compare the results of each of these methods. Machine learning methods applied to the imputed dataset will give better accuracy than that of the incomplete dataset.


Keywords

Missing values, imputation, inverse distance weighted interpolation.

Cite this Article

Arumuga Nainar S., A Comparative Study of Missing Value Imputation Methods on Time Series Data, HCTL Open International Journal of Technology Innovations and Research (IJTIR), Volume 14, April 2015, eISSN: 2321-1814, ISBN (Print): 978-1-62951-946-3.

Download Full-Text PDF

It appears you do not have a PDF plugin for this browser. No issues.. you can click here to download this Full-Text PDF file.

Certificate of Contribution

Certificate of Contribution