Ion channel prediction Using Lightgbm Model
Xinyu Zhang
Abstract
Many diseases, including cancer, is associated with Ion channels. Ion channels are pore-forming proteins in many creatures like animals and plants. They can help to encode memory and learning, fight infections and even enable pain signals. Liverpool university release a dataset of Ion channels to help researchers to find out the relationship between electric signals and ion switching. In this paper, we proposed a lightgbm based model to predict Ion channels numbers. In addition, the data cleaning part and feature engineering part are also included in this paper. Finally, in the experiment section, classical methods are compared with the proposed lightgbm to show the superiority of lightgbm model in both accuary and micro F1 score.