iML challenge

1. Links

• Overleaf
• iML Challenge - results
• Markdown Cheatsheet

2. Methods tested

Linear:

• LinearClassifier : 1.72767
• Rounded LinearRegression : 0.97520
• Rounded LinearRegression with gender added to feature: 0.97520
• Rounded Bagging 10 estimators of linear regression with 0.25 of LS: 0.99946

DT:

• DecisionTreeRegressor : 2.47322 (toy_example)
• DecisionTreeRegressor with gender of user added to feature: 2.45850

RF:

• depth = 8; feat = 0.7; est = 50 : 1.15

kNN

• Bagging 10 estimators of 10-NN with 0.25 of LS (can't more, memory error..): 1.52260
• Rounded Bagging 10 estimators of 10-NN regressor with 0.25 of LS: 1.20338

3. Methods to test

3. Neural network

4. Remarks

• Please use toy_example as template: create output in 'outputs' dir and keep name of your .py file to create the name of the output file.

• don't feel like adding features about users or movies would improve prediction. See tested 4. and 5.