Predicting wind turbine power generation with LightGBM
Our first case study is a problem where we aim to predict the power generation of wind turbines. The dataset for the problem is available from https://www.kaggle.com/datasets/mukund23/hackerearth-machine-learning-challenge.
We work through the problem using the steps defined in the previous section, articulating the details involved in each step alongside code snippets. The complete end-to-end solution is available at https://github.com/PacktPublishing/Practical-Machine-Learning-with-LightGBM-and-Python/tree/main/chapter-6/wind-turbine-power-output.ipynb.
Problem definition
The dataset consists of power generation (in kW/h) measurements of wind turbines taken at a specific date and time. Alongside each measurement are the parameters of the wind turbine, which include physical measurements of the windmill (including windmill height, blade breadth, and length), operating measurements for the turbine (including resistance...
