ML-DSBA-AI-Assignment_2

Assignment proposition for Foundation of Machine Learning courses at CentraleSupélec.

Proposed by Robin Schwob and Paul Asquin.
This assignment is based on the Kaggle contest : www.kaggle.com/c/titanic

Run the models

It's possible to run each model independently bu using the test implemented in test.py. Run the following instructions to execute specific models.
You can also test the load_data function by running:

python3 test.py -m test_load_data

And you can test the load_data_panda function with:

python3 test.py -m test_load_data_panda

Basic classifier

Test the basic classifier given

python3 test.py -m test_basic_classifier

Decision tree

Testing Decision Tree for different depths (best result with D=5)

python3 test.py -m test_decision_tree

Ada boost

Adaboost Test for Different values of D (best with D=2)

python3 test.py -m test_ada_boot

NN

NN Test with sgd, different constant lr, 1 hidden layer of varying size

python3 test.py -m test_NN_1

NN test for higher hidden layer sizes (from 200 to 400)

python3 test.py -m test_NN_2

LDA

python3 test.py -m test_LDA

SVM

python3 test.py -m test_SVM

KNN

python3 test.py -m test_KNN

Random Forest

python3 test.py -m test_random_forest