| Title |
Random Forest-based Training Dataset Supplementation And Prediction Of Lithium-ion Battery Capacity |
| Authors |
Seung-hwa Sin ; Sang-ryuk Lee ; Jong-hoon Kim |
| DOI |
https://doi.org/10.6113/TKPE.2023.28.5.353 |
| Keywords |
Lithium-ion battery; Random forest; ; Regression; Data-driven; LSTM |
| Abstract |
Accurate capacity prediction is essential for the efficient and safe operation of batteries. Recently, data-driven methods that do not require an understanding of the complex degradation mechanisms of batteries have received increasing attention. However, the prediction performance of these methods depends on the quantity and quality of data. In real-life applications, obtaining a sufficient amount of data is difficult due to the irregular operation of batteries. Therefore, this paper proposed a data supplementation method that expands the sample size by predicting the missing data through regression using RF. The permutation importance (PI) automatically calculated during the learning process of RF was used to select important factors and reduce the dimensionality of the feature set. The supplemented dataset using RF was then validated for improved prediction performance using an LSTM model suitable for time-series data. |