In order to explore the effect of missing data on multitask prediction techniques, we assembled complete multitarget data sets to perform activity prediction based on both regression and classification.
In this work, we propose Singular Spectrum Matrix Completion (SS-MC), a novel approach for the simultaneous recovery of missing data and the prediction of future behavior in the absence of complete measurement sets.
Genomic selection models for parent selection were constructed from phenotypic data of their dense-planted half-sib progenies, assessing their selection accuracy for different SNP calling procedures, strategies and algorithms for missing data imputation, and prediction models.
Another way to assess the effect of missing data on the prediction of PA is to examine its effect on the calculation of TEE for a two week period, where TEE was estimated using doubly labeled water (DLW) [ 36].
The effect of allowed missing data thresholds on prediction accuracy, which mostly displayed an accuracy peak in the range of 30 50 %, was consistent with the expected trade-off between increased information (more markers) and increased noise (higher imputation errors) arising from increasing thresholds.
3. Predictive mean matching Available when performing univariate regression imputation or MICE for a continuous variable, the method of predictive mean matching is a partially parametric approach that first predicts the values for the missing data using a linear prediction model.
Severe issues may arise when dealing with missing data for time-series prediction schemes or mean analysis.
Before planning additional experimental trials in order to complete a design matrix, trying to estimate missing data by means of prediction models could be an interesting alternative.
Some efforts have also been dedicated to improve the prediction of missing data, based on correctly received ones (see, e.g., [14, 22] for details).
Hence, this paper focuses on the scenario of a small amount of package losses, and examines a set of estimation methods to mitigate the corrupted and missing data, including substitution, interpolation, and prediction.
