比较python中RF模型的准确性答案

【问题标题】：Compare accuracy of RF model in python比较python中RF模型的准确性
【发布时间】：2018-03-20 16:18:36
【问题描述】：

我想计算它的准确性（在测试数据集上）。该模型具有以下预测值：

[0 1 0 1 1 1 1 0 1 0 1 0 1 1 0 0 0 1 0 1 0 1 0 0 0 1 1 0 0 0 0 0 0 0 1 1 0
 1 1 1 0 0 0 0 0 1 0 0 0 1 0 0 1 0 0 0 0]

如何将其与实际值（在本例中为 B 或 M）进行比较，以获取其对测试数据的准确性。这对于其他数据集值也应该是通用的。这是我用于 RandomForest 模型的代码：

import pandas as pd
import numpy as np
# Load scikit's random forest classifier library
from sklearn.ensemble import RandomForestClassifier

from sklearn.model_selection import train_test_split
file_path = 'https://archive.ics.uci.edu/ml/machine-learning-databases/breast-cancer-wisconsin/wdbc.data'
dataset2 = pd.read_csv(file_path, header=None, sep=',')

train, test = train_test_split(dataset2, test_size=0.1)
y = pd.factorize(train[1])[0]
clf = RandomForestClassifier(n_jobs=2, random_state=0)
features = train.columns[2:]
clf.fit(train[features], y)

RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',
            max_depth=None, max_features='auto', max_leaf_nodes=None,
            min_impurity_split=1e-07, min_samples_leaf=1,
            min_samples_split=2, min_weight_fraction_leaf=0.0,
            n_estimators=10, n_jobs=2, oob_score=False, random_state=0,
            verbose=0, warm_start=False)
# Apply the Classifier we trained to the test data 
clf.predict(test[features])

【问题讨论】：

以下答案是你想做的吗？问题的意思被解释为想用原标签的B、M来评估准确度。

标签： python pandas random-forest

【解决方案1】：

您可以使用 sklearn 的 preprocessing.LabelEncoder() 对 B 和 M 进行编码，如下所示并使用 inverse_transform() 返回。另外，精度评估可以使用pandas_ml包的ConfusionMatrix()和sklearn的accuracy_score()进行。

import pandas as pd
import numpy as np
# Load scikit's random forest classifier library
from sklearn.ensemble import RandomForestClassifier

from sklearn.model_selection import train_test_split
file_path = 'https://archive.ics.uci.edu/ml/machine-learning-databases/breast-cancer-wisconsin/wdbc.data'
dataset2 = pd.read_csv(file_path, header=None, sep=',')

from sklearn import preprocessing
le = preprocessing.LabelEncoder()

# Encode B, M to 0, 1
y = le.fit_transform(dataset2[1])
dataset2[1] = y

train, test = train_test_split(dataset2, test_size=0.1)
y = train[1]
y_test = test[1]
clf = RandomForestClassifier(n_jobs=2, random_state=0)
features = train.columns[2:]
clf.fit(train[features], y)

RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',
            max_depth=None, max_features='auto', max_leaf_nodes=None,
            min_impurity_split=1e-07, min_samples_leaf=1,
            min_samples_split=2, min_weight_fraction_leaf=0.0,
            n_estimators=10, n_jobs=2, oob_score=False, random_state=0,
            verbose=0, warm_start=False)
# Apply the Classifier we trained to the test data
y_pred = clf.predict(test[features])

# Decode from 0, 1 to B, M
y_test_label = le.inverse_transform(y_test)
y_pred_label = le.inverse_transform(y_pred)

from pandas_ml import ConfusionMatrix
confusion_matrix = ConfusionMatrix(y_test_label, y_pred_label)
print("Confusion matrix:\n%s" % confusion_matrix)
# Confusion matrix:
# Predicted   B   M  __all__
# Actual                    
# B          35   1       36
# M           4  17       21
# __all__    39  18       57

from sklearn.metrics import accuracy_score
accuracy_score(y_test_label, y_pred_label)
# Out[14]: 0.035087719298245612

注意pandas_ml可以通过pip轻松安装，如下图所示。

pip install pandas_ml

【讨论】：