ExtraTrees Module

An extra-trees model.

The doc is here : http://scikit-learn.org/dev/modules/generated/sklearn.ensemble.ExtraTreesClassifier.html


Sortie du script

$ python ExtraTrees.py
Opening the file 'train.csv' and 'test.csv'...
Find the best value for the meta parameter n_estimators, with 5 run for each...
Searching in the range : [1, 2, 3, 4, 5, 8, 10, 20, 30, 100]...
Using the first part (66.00%, 588 passengers) of the training dataset as training, 
and the second part (34.00%, 303 passengers) as testing !
For 1 random tree(s), learning from the first part of the dataset...
... this value of n_estimators seems to have a (mean) quality = 87.41%...
For 2 random tree(s), learning from the first part of the dataset...
... this value of n_estimators seems to have a (mean) quality = 88.44%...
For 3 random tree(s), learning from the first part of the dataset...
... this value of n_estimators seems to have a (mean) quality = 87.89%...
For 4 random tree(s), learning from the first part of the dataset...
... this value of n_estimators seems to have a (mean) quality = 88.06%...
For 5 random tree(s), learning from the first part of the dataset...
... this value of n_estimators seems to have a (mean) quality = 88.64%...
For 8 random tree(s), learning from the first part of the dataset...
... this value of n_estimators seems to have a (mean) quality = 87.82%...
For 10 random tree(s), learning from the first part of the dataset...
... this value of n_estimators seems to have a (mean) quality = 88.54%...
For 20 random tree(s), learning from the first part of the dataset...
... this value of n_estimators seems to have a (mean) quality = 88.40%...
For 30 random tree(s), learning from the first part of the dataset...
... this value of n_estimators seems to have a (mean) quality = 88.54%...
For 100 random tree(s), learning from the first part of the dataset...
... this value of n_estimators seems to have a (mean) quality = 88.23%...
With trying each of the following n_estimators ([1, 2, 3, 4, 5, 8, 10, 20, 30, 100]), each 5 times, the best one is 5. (for a quality = 88.64%)
Creating the classifier (of 100 estimators)...
Learning...
 Proportion of perfect fitting for the training dataset = 98.20%
Predicting for the testing dataset
Prediction: wrote in the file csv/ExtraTrees_best.csv.

Résultats

La soumission du résultat à Kaggle donne 75.59%.


ExtraTrees.list_n_estimators = [1, 2, 3, 4, 5, 8, 10, 20, 30, 100]

Espace de recherche

ExtraTrees.Number_try = 5

Nombre de tests utilisés pour méta-apprendre

ExtraTrees.proportion_train = 0.66

Proportion d’individus utilisés pour méta-apprendre.

ExtraTrees.best_n_estimators = 100

La valeur optimale trouvée pour le paramètre n_estimators

ExtraTrees.score = 98.204264870931539

The score for this classifier.

Table des matières

Cette page en .txt et en .pdf

Sujet précédent

DecisionTree Module

Sujet suivant

SVM Module