PCHIP-HW: A Robust Approach for Imputing Microalgae Monitoring Data

Data imputation is crucial for handling missing data in order to ensure the correctness and reliability of the microalgae cultivation monitoring system, particularly for further analysis using the monitoring data. This paper presents a hybrid model called PCHIP-HW, which combines the Piecewise Cubic Hermite Interpolating Polynomial (PCHIP) and Holt-Winters (HW) methods for data imputation. The imputation consists of two stages: the initial stage uses PCHIP, while the latter stage involves decomposition using Holt-Winters, resulting in the generation of season, trend, and level. The second imputation stage involves using PCHIP interpolation to estimate the missing data values for the trend and level of the initial missing data position. Therefore, the interpolation is done based on the missing data index. Afterwards, the imputed data is combined with the season to create the final imputed data. An evaluation was conducted by examining the performance of the PCHIP-HW model under different artificial of the missing value ranges in comparison to the PCHIP approach. The results indicate that PCHIP-HW outperforms traditional PCHIP in all evaluated ranges of missing values. The excellence of PCHIP-HW is evident in its ability to handle empty data and process microalgae data effectively. The outstanding capability enables the generation of complete data for various advanced analysis objectives. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.