DummyClassifier Module

A dummy model, with 3 strategies :

• stratified;
• most_frequent;
• uniform.

The doc is here : http://scikit-learn.org/dev/modules/generated/sklearn.dummy.DummyClassifier.html#sklearn.dummy.DummyClassifier

---

Sortie du script

$ python DummyClassifier.py
Opening the file 'train.csv' and 'test.csv'...
Find the best value for the meta parameter strategy, with 100 run for each...
Searching in the range : ['stratified', 'most_frequent', 'uniform']...
Using the first part (75.00%, 668 passengers) of the training dataset as training, 
and the second part (25.00%, 223 passengers) as testing !
For the strategy stratified, learning from the first part of the dataset...
... this value of strategy seems to have a (mean) quality = 52.34%...
For the strategy most_frequent, learning from the first part of the dataset...
... this value of strategy seems to have a (mean) quality = 60.03%...
For the strategy uniform, learning from the first part of the dataset...
... this value of strategy seems to have a (mean) quality = 50.15%...
With trying each of the following strategy (['stratified', 'most_frequent', 'uniform']), each 100 times, the best one is most_frequent. (for a quality = 60.03%)
Creating the Dummy Classifier classifier with best meta parameters.
Learning...
 Proportion of perfect fitting for the training dataset = 61.62%
Predicting for the testing dataset
Prediction: wrote in the file csv/Dummy_best.csv.

Résultats

La soumission du résultat à Kaggle donne ??.??% (pas encore fait).

---

DummyClassifier.list_strategy = ['stratified', 'most_frequent', 'uniform']

Espace de recherche

DummyClassifier.Number_try = 100

Nombre de tests utilisés pour méta-apprendre

DummyClassifier.proportion_train = 0.75

Proportion d’individus utilisés pour méta-apprendre.

DummyClassifier.best_strategy = 'most_frequent'

La valeur optimale trouvée pour le paramètre strategy

DummyClassifier.score = 61.616161616161612

The score for this classifier.