Missing value imputation affects the performance of machine learning: A review and analysis of the literature (2010–2021)

Missing Value Imputation Designs and Methods of Nature-Inspired Metaheuristic Techniques: A Systematic Review

“…Consequently, a number of literature [10]- [12] discusses recent machine learning-based imputation techniques in solving incomplete dataset problems. Nevertheless, with respect to MVI of nature-inspired metaheuristic techniques, the literature receives limited attention.…”

Challenges of deep learning methods for COVID-19 detection using public datasets

“…We use different metrics, such as recall, precision, F1-score, and accuracy, to evaluate our multi-tasking CVR-Net for COVID-19 recognition, which is mathematically defined [88] as follows: where the TP, FN, FP, and TN respectively denote true positive (patient with coronavirus symptoms recognized as the positive patient), false negative (patient with coronavirus symptoms recognized as the negative patient), false positive (patient without coronavirus symptoms recognized as the positive patient), and true negative (patient without coronavirus symptoms recognized as the negative patient). The recall quantifies the type-II error (the patient, with the positive syndromes, inappropriately fails to be nullified), and precision quantifies the positive predictive values (percentage of truly positive recognition among all the positive recognition).…”

A Novel Missing Data Imputation Approach for Time Series Air Quality Data Based on Logistic Regression

“…The cause for that phenomenon is that the imputation results of FTLRI put forward in this paper only depend on the first five and the last three complete data points, which are highly relevant to the data point with missing values in terms of time and attributes. Instead of relying on all the other complete data points, like other imputation approaches, FTLRI selects the eight data points highly correlated with the missing data point to impute the missing value, which is beneficial for imputation performance [46,47]. Therefore, the increasing of the number of data points and the changing of missing rates will not affect the performance of FTLRI, that is, FTLRI can provide superior imputation results on datasets with different missing rates and different numbers of data points.…”