Using TPOT Classifier to analysis Numerai dataset

numerai.py

"""
Largely come from TPOT example.  The only control you can do to prevent timeout 
and successful running is the "generation" and "population_size" parameters.
Remember, scoring is "log_loss" as of 18 Jan 2017, not probability

The larger the generation and population_size, the longer time you take to get result.
"""

from tpot import TPOTClassifier
from sklearn.model_selection import train_test_split
import pandas as pd

data = pd.read_csv("numerai_training_data.csv")
X, y = data.iloc[:,0:49], data.iloc[:,50]

X_train, X_test, y_train, y_test = train_test_split(X, y, train_size=0.75, test_size=0.25)

tpot = TPOTClassifier(generations=4, population_size=20, verbosity=2, scoring="log_loss")

tpot.fit(X_train, y_train)
print(tpot.score(X_test, y_test))
tpot.export('tpot_numerai_pipeline.py')

tpot_numerai_pipeline-gen-3.py

"""
When using generation = 3
"""
import numpy as np

from sklearn.ensemble import VotingClassifier
from sklearn.model_selection import train_test_split
from sklearn.naive_bayes import MultinomialNB
from sklearn.pipeline import make_pipeline, make_union
from sklearn.preprocessing import FunctionTransformer

# NOTE: Make sure that the class is labeled 'class' in the data file
tpot_data = np.recfromcsv('PATH/TO/DATA/FILE', delimiter='COLUMN_SEPARATOR', dtype=np.float64)
features = np.delete(tpot_data.view(np.float64).reshape(tpot_data.size, -1), tpot_data.dtype.names.index('class'), axis=1)
training_features, testing_features, training_classes, testing_classes = \
    train_test_split(features, tpot_data['class'], random_state=42)

exported_pipeline = make_pipeline(
    MultinomialNB(alpha=0.1, fit_prior=True)
)

exported_pipeline.fit(training_features, training_classes)
results = exported_pipeline.predict(testing_features)

tpot_numerai_pipeline-gen-4.py

"""
When using generation = 4
"""
import numpy as np

from sklearn.ensemble import VotingClassifier
from sklearn.model_selection import train_test_split
from sklearn.naive_bayes import GaussianNB, MultinomialNB
from sklearn.pipeline import make_pipeline, make_union
from sklearn.preprocessing import FunctionTransformer, MinMaxScaler

# NOTE: Make sure that the class is labeled 'class' in the data file
tpot_data = np.recfromcsv('PATH/TO/DATA/FILE', delimiter='COLUMN_SEPARATOR', dtype=np.float64)
features = np.delete(tpot_data.view(np.float64).reshape(tpot_data.size, -1), tpot_data.dtype.names.index('class'), axis=1)
training_features, testing_features, training_classes, testing_classes = \
    train_test_split(features, tpot_data['class'], random_state=42)

exported_pipeline = make_pipeline(
    make_union(
        FunctionTransformer(lambda X: X),
        FunctionTransformer(lambda X: X)
    ),
    MinMaxScaler(),
    make_union(VotingClassifier([("est", GaussianNB())]), FunctionTransformer(lambda X: X)),
    MultinomialNB(alpha=0.71, fit_prior=True)
)

exported_pipeline.fit(training_features, training_classes)
results = exported_pipeline.predict(testing_features)