• Skip to main content
  • Skip to primary sidebar

学習記録

機械学習

機械学習 ワイン

2017年12月7日 by 河副 太智 Leave a Comment

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
# 必要モジュールをインポート
from sklearn.linear_model import LinearRegression
from sklearn.datasets import load_wine
from sklearn.model_selection import train_test_split
 
 
data = load_wine()
# データを教師用とテスト用に
train_X, test_X, train_y, test_y = train_test_split(
    data.data, data.target, random_state=42)
print(data)
# 学習器の構築
model = LinearRegression()
# 教師データを用いて学習器に学習
model.fit(train_X, train_y)
# テスト用データを用いて学習結果をpred_yに代入
pred_y = model.predict(test_X)
# 予測結果を出力
#print(pred_y)

 

Filed Under: 機械学習

機械学習の概念 混同行列

2017年11月27日 by 河副 太智 Leave a Comment

混同行列とは以下の4種類

  1. 真陽性
  2. 偽陽性
  3. 偽陰性
  4. 真陰性

「真か偽」は結果がポジティブかネガティブか
「陽性か陰性」は予想自体が元々ポジティブかネガティブか

 

 

「真」は結果がポジティブとなった

「偽」は結果がネガティブとなった

「陽性」は予想自体が元々ポジティブだった

「陰性」は予想自体が元々ネガティブだった

 

例:1
「医者がエイズであると予測したが実際はエイズではなかった」という場合は
予想の結果は医者の診断に対してネガティブなので「偽」となり、
医者は患者をエイズだ「陽性(ポジティブ)」だと元々予想していたため

このケースでは「偽陽性」となる

 

「エイズと言われエイズだった」場合は真陽性

「エイズと言われエイズで無かった」場合は偽陽性

「エイズではないと言われエイズで無かった」場合は偽陰性

「エイズではないと言われエイズだった」場合は真陰性

 

 

例:2
「友人から結婚は無理と言われて本当に結婚できなかった」場合は
結果は結婚ができなかったので予測は当たりでポジティブ「真」であり
友人の元々の予想は結婚できないと言われネガティブ「陰性」である為

このケースでは「偽陰性」となる

 

「結婚できると言われ結婚できた」場合は真陽性

「結婚できると言われてたが結婚できなかった」場合は偽陽性

「結婚できないと言われ結婚できなかった」場合は偽陰性

「結婚できないと言われ結婚できた」場合は真陰性

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
print(__doc__)
 
import itertools
import numpy as np
import matplotlib.pyplot as plt
 
from sklearn import svm, datasets
from sklearn.model_selection import train_test_split
from sklearn.metrics import confusion_matrix
 
# import some data to play with
iris = datasets.load_iris()
X = iris.data
y = iris.target
class_names = iris.target_names
 
# Split the data into a training set and a test set
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0)
 
# Run classifier, using a model that is too regularized (C too low) to see
# the impact on the results
classifier = svm.SVC(kernel='linear', C=0.01)
y_pred = classifier.fit(X_train, y_train).predict(X_test)
 
 
def plot_confusion_matrix(cm, classes,
                          normalize=False,
                          title='Confusion matrix',
                          cmap=plt.cm.Blues):
    """
    This function prints and plots the confusion matrix.
    Normalization can be applied by setting `normalize=True`.
    """
    if normalize:
        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]
        print("Normalized confusion matrix")
    else:
        print('Confusion matrix, without normalization')
 
    print(cm)
 
    plt.imshow(cm, interpolation='nearest', cmap=cmap)
    plt.title(title)
    plt.colorbar()
    tick_marks = np.arange(len(classes))
    plt.xticks(tick_marks, classes, rotation=45)
    plt.yticks(tick_marks, classes)
 
    fmt = '.2f' if normalize else 'd'
    thresh = cm.max() / 2.
    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):
        plt.text(j, i, format(cm[i, j], fmt),
                 horizontalalignment="center",
                 color="white" if cm[i, j] > thresh else "black")
 
    plt.tight_layout()
    plt.ylabel('True label')
    plt.xlabel('Predicted label')
 
# Compute confusion matrix
cnf_matrix = confusion_matrix(y_test, y_pred)
np.set_printoptions(precision=2)
 
# Plot non-normalized confusion matrix
plt.figure()
plot_confusion_matrix(cnf_matrix, classes=class_names,
                      title='Confusion matrix, without normalization')
 
# Plot normalized confusion matrix
plt.figure()
plot_confusion_matrix(cnf_matrix, classes=class_names, normalize=True,
                      title='Normalized confusion matrix')
 
plt.show()

 

 

Filed Under: 機械学習

機械学習の概念 「一個抜き」(Leave-One-Out:LOO)交差検証

2017年11月27日 by 河副 太智 Leave a Comment

k分割交差検証には、「一個抜き」(Leave-One-Out:LOO)交差検証という方法があり

k分割交差検証から分割の数をデータセットの数と同じにしたもの

(k=データの個数)

 

小さいデータを使用する場合に使われる

Filed Under: 機械学習

機械学習の概念 k分割交差検証

2017年11月27日 by 河副 太智 Leave a Comment

 

k-分割交差検証とはトレーニングデータをk分割(データを”k”個に分ける)
k個から1個引いた個数を学習用に使用し、
引いた1個はテストに使用する

10個に分割すれば9個を学習用1個をテスト用にするという事

これによりk個の学習モデルとk個の性能テストができるので
k回の学習とテストを行い、それらの平均を算出する

これによってホールドアウト法よりバリアンス(過学習)を低く抑えられる

通常kの値は10前後になり、データが多い場合はkの値を増やした方が良い
但し、k値を大きくし過ぎるとバリアンス(過学習)が高くなり、
逆にk値を低くしすぎるとバイアス(学習不足)が高くなる可能性がある

 

Filed Under: 機械学習

機械学習の概念 ホールドアウト法

2017年11月27日 by 河副 太智 Leave a Comment

 

機械学習上の問題に「過学習」(学習し過ぎ)というものがあり
この状態をバリアンスが高いと表現し、
その反対に「学習不足」というものもあり
こちらはバイアスが高いと表現する

 

過学習を防ぐための手段にホールドアウト法というものがある
この方法によって学習データをトレーニングデータとテストデータに分割

トレーニングデータは文言から想像できるように学習という意味で
テストデータというのは学習した内容の評価となる

トレーニングデータという概念は更に分割され、
トレーニングデータサブセットと
検証データセットという2つになります。

これはテストデータがトレーニングデータの一部になってしまう事を防ぐ為にある

 

トレーニングデータサブセットと検証データセットの2つで学習を行い
もう一方のテストデータを用いて最終テストを行う

 

このホールドアウト法の他に
「k分割交差検証」と呼ばれるものもある

Filed Under: 機械学習

  • « Go to Previous Page
  • Page 1
  • Interim pages omitted …
  • Page 7
  • Page 8
  • Page 9

Primary Sidebar

カテゴリー

  • AWS
  • Bootstrap
  • Dash
  • Django
  • flask
  • GIT(sourcetree)
  • Plotly/Dash
  • VPS
  • その他tool
  • ブログ
  • プログラミング
    • Bokeh
    • css
    • HoloViews
    • Jupyter
    • Numpy
    • Pandas
    • PosgreSQL
    • Python 基本
    • python3
      • webアプリ
    • python3解説
    • scikit-learn
    • scipy
    • vps
    • Wordpress
    • グラフ
    • コマンド
    • スクレイピング
    • チートシート
    • データクレンジング
    • ブロックチェーン
    • 作成実績
    • 時系列分析
    • 機械学習
      • 分析手法
      • 教師有り
    • 異常値検知
    • 自然言語処理
  • 一太郎
  • 数学
    • sympy
      • 対数関数(log)
      • 累乗根(n乗根)
    • 暗号学

Copyright © 2025 · Genesis Sample on Genesis Framework · WordPress · Log in