Skip to content

Latest commit

 

History

History
19 lines (14 loc) · 1010 Bytes

File metadata and controls

19 lines (14 loc) · 1010 Bytes

KDD Cup 2022 - Baidu Spatial Dynamic Wind Power Forecasting

This is the solution from team trymore, winning 2th place in 2490 teams.

0. Overview

Based on the weighted hybrid prediction method, our method designs three main modules, each of which plays a different role in the final prediction. In order to obtain better integration results, we used different feature sets for different models to build differences. Next, we will introduce our data processing and the details of the three component models.

model

1. Dataset

You can download the data 'wtbdata_245days.csv' from Baidu AI studio, https://aistudio.baidu.com/aistudio/competition/detail/152/0/datasets.

2. Train

Details can be find in ./train_code/Readme.md, fellow the steps you will get all the models.

3. Inference

The inference code is located in ./test_code.