ABSTRACT
Missing values in hydrological studies are a common issue for hydrologists, especially in statistical analyses as a complete dataset is required. This work evaluates the performance of the multiple imputations by chained equations (MICE) approach to predicting recurrence in streamflow datasets. To evaluate and verify the effectiveness of the MICE approach in treating missing streamflow data, complete historical daily streamflow data from 2012 to 2014 were used. Later, MICE methods coupled with multiple linear regression (MLR) were applied to restore streamflow rates in Malaysia’s Langat River basin from 1978 to 2016. The best estimation methods are validated with tests such as adjusted R-squared (Adj R2), residual standard error (RSE), and mean absolute percentage error (MAPE). The findings revealed that the classification and regression tree (CART) method combined with MLR outperformed the other approaches tested, with the highest Adj R2 value and the lowest RSE and MAPE values observed regardless of missing conditions.
EDITOR:
ASSOCIATE EDITOR:
Disclosure statement
No potential conflict of interest was reported by the authors.