Extreme learning machine network with TSK fuzzy type

摘要

This study suggests an ELM (Extreme Learning Machine) model that is based on a TSK (Takagi-Sugeno-Kang) fuzzy model and compares its performance projection with the existing ELM model. The TSK based ELM model replaces the in the existing model with a linear function. Additionally, the center of the cluster is haphazardly set. The Weighted value between the hidden layer and input is nonexistent whereas the weighted value between the hidden layer and output apply a linear equation. Using the data from the short-term power load data, which The performance predictions of the two models comparing power load predictions were compared. In conclusion, the TSK based ELM model outperformed the current ELM model in performance projection.