Árboles de Decisión
Datos disponibles en https://www.kaggle.com/c/titanic/data?select=train.csv
Un árbol de decisión es una forma gráfica y analítica que presenta sucesos y sus posibles consecuencias.
Ventajas:
Claridad en los datos
Tolerantes al ruido y valores faltantes
Las reglas extraídas permiten hacer predicciones
Desventajas:
Criterio de división es deficiente
Sobreajuste
Ramas poco significativas
Cómo se divide un árbol de decisión?
Ganancia de Información
Dividir en pequeños árboles
Optimización del modelo:
Evitar sobreajuste
Selección de atributos, evaluando cuales son aquellos atributos que nos permitirán dividir mejor y ganar más información al hacerlo.
No tener campos nulos, ya que estos agregan mucho ruido al árbol y afecta a la división de la información
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import re
import numpy as np
from sklearn import tree
from sklearn.model_selection import train_test_split
%matplotlib inline
sns.set()
test_df = pd.read_csv('/work/titanic-test.csv')
train_df = pd.read_csv('/work/titanic-train.csv')
train_df.head()
PassengerIdint64
Survivedint64
0
1
0
1
2
1
2
3
1
3
4
1
4
5
0
train_df.shape
test_df.shape
train_df.info()
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 891 entries, 0 to 890
Data columns (total 12 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 PassengerId 891 non-null int64
1 Survived 891 non-null int64
2 Pclass 891 non-null int64
3 Name 891 non-null object
4 Sex 891 non-null object
5 Age 714 non-null float64
6 SibSp 891 non-null int64
7 Parch 891 non-null int64
8 Ticket 891 non-null object
9 Fare 891 non-null float64
10 Cabin 204 non-null object
11 Embarked 889 non-null object
dtypes: float64(2), int64(5), object(5)
memory usage: 83.7+ KB
test_df.info()
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 418 entries, 0 to 417
Data columns (total 11 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 PassengerId 418 non-null int64
1 Pclass 418 non-null int64
2 Name 418 non-null object
3 Sex 418 non-null object
4 Age 332 non-null float64
5 SibSp 418 non-null int64
6 Parch 418 non-null int64
7 Ticket 418 non-null object
8 Fare 417 non-null float64
9 Cabin 91 non-null object
10 Embarked 418 non-null object
dtypes: float64(2), int64(4), object(5)
memory usage: 36.0+ KB
train_df.Sex.value_counts().plot(kind = 'bar', color = ['b', 'r'])
plt.title('Distribución de pasajeros por Género')
plt.show()
train_df[ train_df['Survived'] == 1 ]['Sex'].value_counts().plot(kind='bar', color=['b','r'])
plt.title('Distribucion de sobrevivientes')
from sklearn import preprocessing
label_encoder = preprocessing.LabelEncoder()
encoder_sex = label_encoder.fit_transform(train_df['Sex'])
train_df.head()
PassengerIdint64
Survivedint64
0
1
0
1
2
1
2
3
1
3
4
1
4
5
0
train_df.isnull().any()
test_df.isnull().any()
train_df['Age'] = train_df['Age'].fillna(train_df['Age'].median())
train_df['Embarked'] = train_df['Embarked'].fillna('S')
train_predictors = train_df.drop(['PassengerId', 'Survived', 'Name', 'Ticket', 'Cabin'], axis = 1)
categorical_cols = [cname for cname in train_predictors.columns if
train_predictors[cname].nunique() < 10 and
train_predictors[cname].dtype == 'object'
]
numerical_cols = [cname for cname in train_predictors.columns if
train_predictors[cname].dtype in ['int64', 'float64']
]
my_cols = categorical_cols + numerical_cols
train_predictors = train_predictors[my_cols]
dummy_encoded_train_predictors = pd.get_dummies(train_predictors)
train_df['Pclass'].value_counts()
x_features_one = dummy_encoded_train_predictors.values
y_target = train_df['Survived'].values
x_train, x_validation, y_train, y_validation = train_test_split(x_features_one, y_target, test_size = .25, random_state = 1)
tree_one = tree.DecisionTreeClassifier()
tree_one = tree_one.fit(x_features_one, y_target)
tree_one_accuracy = round(tree_one.score(x_features_one, y_target), 4)
print('Accuracy: %0.4f' % (tree_one_accuracy))
Accuracy: 0.9798
! pip install pydotplus
Collecting pydotplus
Downloading pydotplus-2.0.2.tar.gz (278 kB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 278.7/278.7 KB 5.3 MB/s eta 0:00:00
Preparing metadata (setup.py) ... done
Requirement already satisfied: pyparsing>=2.0.1 in /shared-libs/python3.9/py-core/lib/python3.9/site-packages (from pydotplus) (3.0.9)
Building wheels for collected packages: pydotplus
Building wheel for pydotplus (setup.py) ... done
Created wheel for pydotplus: filename=pydotplus-2.0.2-py3-none-any.whl size=24575 sha256=a8bf3b56c3f87f58ca4c03c9fe96543b5256d5953570b9487efad18857d0acba
Stored in directory: /root/.cache/pip/wheels/89/e5/de/6966007cf223872eedfbebbe0e074534e72e9128c8fd4b55eb
Successfully built pydotplus
Installing collected packages: pydotplus
Successfully installed pydotplus-2.0.2
WARNING: You are using pip version 22.0.4; however, version 22.3.1 is available.
You should consider upgrading via the '/root/venv/bin/python -m pip install --upgrade pip' command.
from io import StringIO
from IPython.display import Image, display
import pydotplus
out = StringIO()
tree.export_graphviz(tree_one, out_file = out)
graph = pydotplus.graph_from_dot_data(out.getvalue())
graph.write_png('titanic.png')