PRACTICE Beginner Challenge
Can you predict a measure of wealth for different locations across Africa?
Ended ~1 year ago
107 active · 553 enrolled
Good for beginners
Financial Services
The right algorithm ?
Data · 13 Mar 2022, 17:43 · 5

which algorithm should i use ? i used RandomForestRegressor

Discussion 5 answers

honestly I'm not sure. but in the jupyter starter notebook they recommended using xgboost and neural networks saying it's better than randomforest in this task.

they also used catboostregressor at the end.

13 Mar 2022, 18:42
Upvotes 0

Try go with XGBoost, CatBoost and LGBM. They would give you the best results.

Just be sure to tune the hyperparameters properly.

Also, consider using the frequency encoder for the categorical variables.

I tried xgboost but it froze my computer for some reason.. both on jupyter and on pycharm :(

That's weird.

Try put the hyperparameter

n_jobs = 1

In the definition of xgboost

lgbm, catboost , xgboost ,gradient boosting regressor all did well . i'm yet to properly tune the hyperparameters so you can start with these

14 Mar 2022, 15:52
Upvotes 0