CallMeAmartya / Food-Demand-Forecasting

hypertuning decision tree to make predictions

Genpact Machine Learning Hackathone:Food-Demand-Forecasting

The model based on Decision Tree algorithm predicts the number of food orders for various cities for a meal delivery company.

Decision Tree Model trained on train.csv The evaluation metric used is 100*RMSLE where RMSLE is Root of Mean Squared Logarithmic Error across all entries in the test set. Feature selection using Pearson correlation matrix. Hyperparametric tuning isnt done.

The code can be run on jupyter notebook through anaconda prompt. The model learned then make predictions on entries in test.csv and automatically publishes a csv file containing id (test id) and relative prediction.