/predicting-crop-yield

Applying various handmade ML algorithms to predict worldwide crop yields.

Primary LanguageJupyter Notebook

Predicting worldwide crop yield

Applying various handmade ML algorithms to predict worldwide crop yields

• Performed exploratory analysis, data processing, PCA and feature engineering on worldwide weather and crop yield data from 1980-present (35K observations), splitting into train, validation and test sets.

• Implemented from scratch and analysed the performance of multiple regression algorithms on the prediction of crop yield. (Linear regression, Random forest, K-NN and Gaussian Processes).

• Tuned all hyper-parameters via cross-validation and analysed the final performance RMSE & R-squared metrics.

My final report is visible under "ML1 REPORT.pdf"

Final grade: First class (94%)