学習記録 – ページ 28

ランダムフォレスト

2018年1月16日 by 河副太智 Leave a Comment

ランダムフォレスト

train_X = df.drop('Survived', axis=1) train_y = df.Survived (
train_X, test_X ,train_y, test_y
) = train_test_split(train_X, train_y, test_size = 0.3, random_state = 666)


#ランダムフォレスト
from sklearn.ensemble import RandomForestClassifier
clf = RandomForestClassifier(random_state=0)
clf = clf.fit(train_X, train_y)
pred = clf.predict(test_X)
fpr, tpr, thresholds = roc_curve(test_y, pred, pos_label=1)
auc(fpr, tpr)
accuracy_score(pred, test_y)

train_X = df.drop('Survived', axis=1) train_y = df.Survived (

train_X, test_X ,train_y, test_y

) = train_test_split(train_X, train_y, test_size = 0.3, random_state = 666)

#ランダムフォレスト

from sklearn.ensemble import RandomForestClassifier

clf = RandomForestClassifier(random_state=0)

clf = clf.fit(train_X, train_y)

pred = clf.predict(test_X)

fpr, tpr, thresholds = roc_curve(test_y, pred, pos_label=1)

auc(fpr, tpr)

accuracy_score(pred, test_y)

criterion : 分割基準。gini or entropyを選択。(デフォルトでジニ係数)
max_depth : 木の深さ。木が深くなるほど過学習し易いので、適当なしきい値を設定してあげる。
max_features：最適な分割をする際の特徴量の数
min_samples_split：分岐する際のサンプル数
random_state：ランダムseedの設定。seedを設定しないと、毎回モデル結果が変わるので注意。

公式ドキュメント
sklearn.tree.DecisionTreeClassifier — scikit-learn 0.19.1 documentation

どの要素が強く影響を与えているかを確認

import matplotlib.pyplot as plt
import numpy as np
%matplotlib inline

features = train_X.columns
importances = clf.feature_importances_
indices = np.argsort(importances)

plt.figure(figsize=(6,6))
plt.barh(range(len(indices)), importances[indices], color='b', align='center')
plt.yticks(range(len(indices)), features[indices])
plt.show()

import matplotlib.pyplot as plt

import numpy as np

%matplotlib inline

features = train_X.columns

importances = clf.feature_importances_

indices = np.argsort(importances)

plt.figure(figsize=(6,6))

plt.barh(range(len(indices)), importances[indices], color='b', align='center')

plt.yticks(range(len(indices)), features[indices])

plt.show()

決定木の可視化　ツリーグラフ

2018年1月16日 by 河副太智 Leave a Comment

最初にインストール

pip install pydotplus
brew install graphviz

1 2	pip install pydotplus brew install graphviz

#可視化
import pydotplus
from IPython.display import Image
from graphviz import Digraph
from sklearn.externals.six import StringIO

dot_data = StringIO()
tree.export_graphviz(clf, out_file=dot_data,feature_names=train_X.columns, max_depth=3)
graph = pydotplus.graph_from_dot_data(dot_data.getvalue())
graph.write_pdf("graph.pdf")
Image(graph.create_png())

#可視化

import pydotplus

from IPython.display import Image

from graphviz import Digraph

from sklearn.externals.six import StringIO

dot_data = StringIO()

tree.export_graphviz(clf, out_file=dot_data,feature_names=train_X.columns, max_depth=3)

graph = pydotplus.graph_from_dot_data(dot_data.getvalue())

graph.write_pdf("graph.pdf")

Image(graph.create_png())

aucモデル評価、モデルスコア

2018年1月16日 by 河副太智 Leave a Comment

モデルスコア1に近いほど精度が高い

from sklearn.metrics import (roc_curve, auc, accuracy_score)

pred = clf.predict(test_X)
fpr, tpr, thresholds = roc_curve(test_y, pred, pos_label=1)
auc(fpr, tpr)
accuracy_score(pred, test_y)

from sklearn.metrics import (roc_curve, auc, accuracy_score)

pred = clf.predict(test_X)

fpr, tpr, thresholds = roc_curve(test_y, pred, pos_label=1)

auc(fpr, tpr)

accuracy_score(pred, test_y)

データのインポート

2018年1月16日 by 河副太智 Leave a Comment

インポート用コード

pd.read_csv(filename) # From a CSV file
pd.read_table(filename) # From a delimited text file (like TSV)
pd.read_excel(filename) # From an Excel file
pd.read_sql(query, connection_object) # Reads from a SQL table/database
pd.read_json(json_string) # Reads from a JSON formatted string, URL or file.
pd.read_html(url) # Parses an html URL, string or file and extracts tables to a list of dataframes
pd.read_clipboard() # Takes the contents of your clipboard and passes it to read_table()
pd.DataFrame(dict) # From a dict, keys for columns names, values for data as lists

pd.read_csv(filename) # From a CSV file

pd.read_table(filename) # From a delimited text file (like TSV)

pd.read_excel(filename) # From an Excel file

pd.read_sql(query, connection_object) # Reads from a SQL table/database

pd.read_json(json_string) # Reads from a JSON formatted string, URL or file.

pd.read_html(url) # Parses an html URL, string or file and extracts tables to a list of dataframes

pd.read_clipboard() # Takes the contents of your clipboard and passes it to read_table()

pd.DataFrame(dict) # From a dict, keys for columns names, values for data as lists

データの選択、特定

2018年1月16日 by 河副太智 Leave a Comment

選択、特定

df[col] # Returns column with label col as Series
df[[col1, col2]] # Returns Columns as a new DataFrame
s.iloc[0] # Selection by position (selects first element)
s.loc[0] # Selection by index (selects element at index 0)
df.iloc[0,:] # First row
df.iloc[0,0] # First element of first column

df[col] # Returns column with label col as Series

df[[col1, col2]] # Returns Columns as a new DataFrame

s.iloc[0] # Selection by position (selects first element)

s.loc[0] # Selection by index (selects element at index 0)

df.iloc[0,:] # First row

df.iloc[0,0] # First element of first column

データクリーニング

2018年1月16日 by 河副太智 Leave a Comment

データクリーニング、データクレンジング

Data CleaningPython

df.columns = ['a','b','c'] # Renames columns
pd.isnull() # Checks for null Values, Returns Boolean Array
pd.notnull() # Opposite of s.isnull()
df.dropna() # Drops all rows that contain null values
df.dropna(axis=1) # Drops all columns that contain null values
df.dropna(axis=1,thresh=n) # Drops all rows have have less than n non null values
df.fillna(x) # Replaces all null values with x
s.fillna(s.mean()) # Replaces all null values with the mean (mean can be replaced with almost any function from the statistics section)
s.astype(float) # Converts the datatype of the series to float
s.replace(1,'one') # Replaces all values equal to 1 with 'one'
s.replace([1,3],['one','three']) # Replaces all 1 with 'one' and 3 with 'three'
df.rename(columns=lambda x: x + 1) # Mass renaming of columns
df.rename(columns={'old_name': 'new_ name'}) # Selective renaming
df.set_index('column_one') # Changes the index
df.rename(index=lambda x: x + 1) # Mass renaming of index

Data CleaningPython

df.columns = ['a','b','c'] # Renames columns

pd.isnull() # Checks for null Values, Returns Boolean Array

pd.notnull() # Opposite of s.isnull()

df.dropna() # Drops all rows that contain null values

df.dropna(axis=1) # Drops all columns that contain null values

df.dropna(axis=1,thresh=n) # Drops all rows have have less than n non null values

df.fillna(x) # Replaces all null values with x

s.fillna(s.mean()) # Replaces all null values with the mean (mean can be replaced with almost any function from the statistics section)

s.astype(float) # Converts the datatype of the series to float

s.replace(1,'one') # Replaces all values equal to 1 with 'one'

s.replace([1,3],['one','three']) # Replaces all 1 with 'one' and 3 with 'three'

df.rename(columns=lambda x: x + 1) # Mass renaming of columns

df.rename(columns={'old_name': 'new_ name'}) # Selective renaming

df.set_index('column_one') # Changes the index

df.rename(index=lambda x: x + 1) # Mass renaming of index