Comparative Performance Analysis of Filling Missing Values Algorithms in PdM Systems of UAV
Abstract
With the development of the IoT domain, the volume of data produced by various applications has also increased. Due to multiple reasons, such as sensor failure, communication system failure, and human errors, the data acquired from the sensors have missing values. The presence of missing values in the dataset affects the informational content of the dataset and thus affects the process of extracting knowledge from the data. In this paper, the authors present a comparative analysis of the performances of the methods of filling in the missing values, such as method, Interpolation, Mean, the K-Nearest Neighbors (KNN), and Random Forests (RF), on the data coming from a Predictive Maintenance (PdM) system that can be used at Unmanned Aerial Vehicle (UAV). The data on which the performance of these methods has been studied comes from a PdM system from the UAVs, used to identify the defects of the Brushless DC (BLDC) motors and estimate the Remaining Useful Life (RUL) of Li-ion batteries.