文章基本信息

标题：Evaluating the Performance of Multiple Imputation Methods for Handling Missing Values in Time Series Data: A Study Focused on East Africa, Soil-Carbonate-Stable Isotope Data
本地全文：下载
作者：Hossein Hassani ; Mahdi Kalantari ; Zara Ghodsi 等
期刊名称：Stats
电子版ISSN：2571-905X
出版年度：2019
卷号：2
期号：4
页码：457-467
DOI：10.3390/stats2040032
出版社：MDPI AG
摘要：In all fields of quantitative research, analysing data with missing values is an excruciating challenge. It should be no surprise that given the fragmentary nature of fossil records, the presence of missing values in geographical databases is unavoidable. As in such studies ignoring missing values may result in biased estimations or invalid conclusions, adopting a reliable imputation method should be regarded as an essential consideration. In this study, the performance of singular spectrum analysis (SSA) based on L 1 norm was evaluated on the compiled δ 13 C data from East Africa soil carbonates, which is a world targeted historical geology data set. Results were compared with ten traditionally well-known imputation methods showing L 1 -SSA performs well in keeping the variability of the time series and providing estimations which are less affected by extreme values, suggesting the method introduced here deserves further consideration in practice.
关键词：imputation; soil carbonate; missing values; fossil records; Africa imputation ; soil carbonate ; missing values ; fossil records ; Africa