Scikit-learn机器学习算法库代码实践

关注

原创最新推荐文章于 2026-04-26 09:06:30 发布 568 阅读 AI 写同款· GEO 优化 ›

（一）由源码学算法：通过scikit-learn库代码示例学机器学习算法（总）通过scikit-learn库代码示例学机器学习算法（总）对于刚转行或新入门机器学习的同志，算法作为代码的骨架，是最核心、最坚挺的部分，无疑也是最需要关切、最紧迫的事情。一般学习算法的思路是老师先讲授数据结构，然后讲解算法，再通过代码示例说明，这样做的好处是显而易见的，算法可以学的更通透。然而，如果想要在短期内能够实现代码能力codingCapicity则显得冗余耗时，算法精了... 阅读详情

Scikit-learn机器学习库包含的机器学习方法

官网直达scikit-learn: machine learning in Python — scikit-learn 0.21.3 documentation
https://scikit-learn.org/stable/

在这里插入图片描述

本例中用到的方法有：

    classifiers = {'NB':naive_bayes_classifier, 
                  'KNN':knn_classifier,
                   'LR':logistic_regression_classifier,
                   'RF':random_forest_classifier,
                   'DT':decision_tree_classifier,
                  'SVM':svm_classifier,
                'SVMCV':svm_cross_validation,
                 'GBDT':gradient_boosting_classifier
    }

1 加载训练集和测试集数据

核心代码注解

# 使用gizp打开文件，read bytes字节方式读取
f = gzip.open(data_file,"rb")
# 使用pickle方法从序列化的数据中按照字节方式加载到内存
train, val, test = pickcle.load(f, encoding = 'bytes')
# 获取训练数据集的样本数和特征维度
num_train, num_feat = train_x.shape
# 对训练集的标记进行去重处理，判断输出空间的维度，如果是2，就说明是一个二分类问题
is_binary_class = (len(np.unique(train_y)) == 2)

Python数据存储：pickle模块的使用讲解 - 风-fmgao - 博客园
https://www.cnblogs.com/fmgao-technology/p/9078918.html

完整代码如下

# python中对数据进行序列化，并保存到对象中的操作模块
import pickle
# 原本是linux中的解压缩命令
import gzip
# 开始读取数据
print('reading training and testing data...') 
# 使用gzip里的方法打开数据文件
f = gzip.open(data_file, "rb")
# 编码方式用bytes字节而不是ascll字符，也不是utf-8，这里自动就加载了训练集和测试集
train, val, test = pickle.load(f,encoding='bytes')
f.close()
# 第一列是特征
train_x = train[0]
# 第二列是标记
train_y = train[1]
test_x = test[0]
test_y = test[1]
# 获取大小，结果是(50000, 784)，即50000行，784列，784表示特征的维度
num_train, num_feat = train_x.shape
# 大小是(10000, 784)
num_test, num_feat = test_x.shape
# np.unique(train_y)表示测试数据集去重之后的结果，测试数据集包含0~9的数据
# 如果这个长度是2，说明数字只有两类，变成二分类问题
is_binary_class = (len(np.unique(train_y)) == 2)
print('******************** Data Info *********************') 
print('#training data: %d, #testing_data: %d, dimension: %d' % (num_train, num_test, num_feat))

输出结果：

reading training and testing data...
******************** Data Info *********************
#training data: 50000, #testing_data: 10000, dimension: 784

2 模型训练，预测和性能测试报告

关键代码分析

# 统计花费的时间
time_cost = time.time() - start_time
# 选择模型
model = MultinomialNB(alpha=0.01)
# 模型训练（在训练集上）
model.fit(train_x, train_y)
# 模型预测（在测试集上）
predict = model.predict(test_x)
# 模型性能报告
accuracy = metrics.accuracy_score(test_y, predict)

metrics是机器学习方法的性能指标

机器学习常用性能指标总结 - Prince1994 - 博客园
https://www.cnblogs.com/princecoding/p/6714216.html

scikit-learn 朴素贝叶斯类库使用小结 - 刘建平Pinard - 博客园
https://www.cnblogs.com/pinard/p/6074222.html

以朴素贝叶斯为例，代码如下

# 开始计时
start_time = time.time()
# 使用一个朴素贝叶斯分类器训练模型
# 从机器学习库中的朴素贝叶斯加载多分类朴素贝叶斯模型
from sklearn.naive_bayes import MultinomialNB
# 选择模型，设置学习率
model = MultinomialNB(alpha=0.01)
# 开始训练
model.fit(train_x, train_y)
# 打印训练模型花费的时间
print('training took %fs!' % (time.time() - start_time)) 
# 对测试集进行预测
predict = model.predict(test_x)
# 如果是二分类问题，还需要计算准确率和召回率
if is_binary_class:
    precision = metrics.precision_score(test_y, predict)
    recall = metrics.recall_score(test_y, predict)
    print('precision: %.2f%%, recall: %.2f%%' % (100 * precision, 100 * recall)) 
# 计算模型在测试集上的准确性
accuracy = metrics.accuracy_score(test_y, predict)
print('accuracy: %.2f%%' % (100 * accuracy) )

输出结果

training took 0.244317s!
accuracy: 83.69%

从python2到python3需要注意的地方

参考的原代码是python2写的，自己在编译的时候遇到了很多坑，总结如下：

print需要改为print()
import cPickle需要改为import pickle
import sys改成import importlib
reload(sys)改成importlib.reload(sys)
sys.setdefaultencoding('utf8')直接去掉，因为utf8是python3的默认编码格式
train, val, test = pickle.load(f)需要改成train, val, test = pickle.load(f,encoding='bytes')

踩坑避难参考成功解决Python3版UnicodeDecodeError: ‘ascii’ codec can’t decode byte 0x90 in position 614: ordinal not in - 一个处女座的程序猿 - 优快云博客

完整代码附录和运行结果

#!usr/bin/env python
#-*- coding: utf-8 -*-
# python2的加载 
# import sys
# import cPickle as pickle
# reload(sys)
# sys.setdefaultencoding('utf8')

import os
# 统计时间
import time
# 性能度量
from sklearn import metrics
import numpy as np
# python数据存储模块
import pickle
import importlib,sys

importlib.reload(sys)

# Multinomial Naive Bayes Classifier
def naive_bayes_classifier(train_x, train_y):
    from sklearn.naive_bayes import MultinomialNB
    model = MultinomialNB(alpha=0.01)
    model.fit(train_x, train_y)
    return model
 
 
# KNN Classifier
def knn_classifier(train_x, train_y):
    from sklearn.neighbors import KNeighborsClassifier
    model = KNeighborsClassifier()
    model.fit(train_x, train_y)
    return model
 
 
# Logistic Regression Classifier
def logistic_regression_classifier(train_x, train_y):
    from sklearn.linear_model import LogisticRegression
    model = LogisticRegression(penalty='l2')
    model.fit(train_x, train_y)
    return model
 
 
# Random Forest Classifier
def random_forest_classifier(train_x, train_y):
    from sklearn.ensemble import RandomForestClassifier
    model = RandomForestClassifier(n_estimators=8)
    model.fit(train_x, train_y)
    return model
 
 
# Decision Tree Classifier
def decision_tree_classifier(train_x, train_y):
    from sklearn import tree
    model = tree.DecisionTreeClassifier()
    model.fit(train_x, train_y)
    return model
 
 
# GBDT(Gradient Boosting Decision Tree) Classifier
def gradient_boosting_classifier(train_x, train_y):
    from sklearn.ensemble import GradientBoostingClassifier
    model = GradientBoostingClassifier(n_estimators=200)
    model.fit(train_x, train_y)
    return model
 
 
# SVM Classifier
def svm_classifier(train_x, train_y):
    from sklearn.svm import SVC
    model = SVC(kernel='rbf', probability=True)
    model.fit(train_x, train_y)
    return model
 
# SVM Classifier using cross validation
def svm_cross_validation(train_x, train_y):
    from sklearn.grid_search import GridSearchCV
    from sklearn.svm import SVC
    model = SVC(kernel='rbf', probability=True)
    param_grid = {'C': [1e-3, 1e-2, 1e-1, 1, 10, 100, 1000], 'gamma': [0.001, 0.0001]}
    grid_search = GridSearchCV(model, param_grid, n_jobs = 1, verbose=1)
    grid_search.fit(train_x, train_y)
    best_parameters = grid_search.best_estimator_.get_params()
    for para, val in best_parameters.items():
        print(para, val)
    model = SVC(kernel='rbf', C=best_parameters['C'], gamma=best_parameters['gamma'], probability=True)
    model.fit(train_x, train_y)
    return model
 
def read_data(data_file):
    # 加载一个能解压缩的库
    import gzip
    # 使用gzip里的方法打开数据文件
    f = gzip.open(data_file, "rb")
#     train, val, test = pickle.load(f)
	# 编码方式用bytes，这里自动就加载了训练集和测试集
    train, val, test = pickle.load(f,encoding='bytes')
    f.close()
    train_x = train[0]
    train_y = train[1]
    test_x = test[0]
    test_y = test[1]
    # 返回训练集和测试集
    return train_x, train_y, test_x, test_y
    
if __name__ == '__main__':
    # 加载数据集，注意是未解压缩的
    data_file = "mnist.pkl.gz"
    thresh = 0.5
    model_save_file = None
    model_save = {}
    
    test_classifiers = ['NB', 'KNN', 'LR', 'RF', 'DT', 'SVM', 'GBDT']
    classifiers = {'NB':naive_bayes_classifier, 
                  'KNN':knn_classifier,
                   'LR':logistic_regression_classifier,
                   'RF':random_forest_classifier,
                   'DT':decision_tree_classifier,
                  'SVM':svm_classifier,
                'SVMCV':svm_cross_validation,
                 'GBDT':gradient_boosting_classifier
    }
    
    print('reading training and testing data...') 
    train_x, train_y, test_x, test_y = read_data(data_file)
    num_train, num_feat = train_x.shape
    num_test, num_feat = test_x.shape
    is_binary_class = (len(np.unique(train_y)) == 2)
    print('******************** Data Info *********************') 
    print('#training data: %d, #testing_data: %d, dimension: %d' % (num_train, num_test, num_feat)) 
    
    # 对每一个分类器进行遍历
    for classifier in test_classifiers:
        print('******************* %s ********************' % classifier) 
        # 开始计时
        start_time = time.time()
        # 使用分类器训练模型
        model = classifiers[classifier](train_x, train_y)
        # 打印训练模型花费的时间
        print('training took %fs!' % (time.time() - start_time)) 
        # 对测试集进行预测
        predict = model.predict(test_x)
        if model_save_file != None:
            model_save[classifier] = model
        # 如果是二分类问题，还需要计算准确率和召回率
        if is_binary_class:
            precision = metrics.precision_score(test_y, predict)
            recall = metrics.recall_score(test_y, predict)
            print('precision: %.2f%%, recall: %.2f%%' % (100 * precision, 100 * recall)) 
        # 计算准确性
        accuracy = metrics.accuracy_score(test_y, predict)
        print('accuracy: %.2f%%' % (100 * accuracy) ) 
 
    if model_save_file != None:
        pickle.dump(model_save, open(model_save_file, 'wb'))

本次使用mnist手写体库进行实验：http://deeplearning.net/data/mnist/mnist.pkl.gz，共5万训练样本和1万测试样本。

运行效果：

reading training and testing data...
******************** Data Info *********************
#training data: 50000, #testing_data: 10000, dimension: 784
******************* NB ********************
training took 0.295175s!
accuracy: 83.69%
******************* KNN ********************
training took 15.932032s!
accuracy: 96.64%
******************* LR ********************
training took 52.029990s!
accuracy: 91.99%
******************* RF ********************
training took 2.521217s!
accuracy: 93.85%
******************* DT ********************
training took 16.277198s!
accuracy: 87.45%
******************* SVM ********************
training took 2627.654293s!
accuracy: 94.35%
******************* GBDT ********************
training took 4673.062296s!
accuracy: 96.23%