机器学习、数据挖掘和统计模式识别学习（Matlab代码实现）

       目录

&＃x1f4a5;1 概述

&＃x1f4da;2 运行结果

&＃x1f389;3 参考文献

&＃x1f468;‍&＃x1f4bb;4 Matlab代码

&＃x1f4a5;1 概述

机器学习是让计算机在没有明确编程的情况下采取行动的科学。在过去的十年中&＃xff0c;机器学习为我们提供了自动驾驶汽车&＃xff0c;实用的语音识别&＃xff0c;有效的网络搜索以及对人类基因组的理解大大提高。机器学习在今天是如此普遍&＃xff0c;以至于你可能每天使用它几十次而不自知。许多研究人员还认为&＃xff0c;这是朝着人类水平的人工智能取得进展的最佳方式。在本代码中&＃xff0c;您将了解最有效的机器学习技术&＃xff0c;并获得实施它们并让它们为自己工作的练习。更重要的是&＃xff0c;您不仅将学习学习的理论基础&＃xff0c;还将获得快速有效地将这些技术应用于新问题所需的实践知识。最后&＃xff0c;您将了解硅谷在创新方面的一些最佳实践&＃xff0c;因为它与机器学习和人工智能有关。本代码广泛介绍了机器学习、数据挖掘和统计模式识别。主题包括&＃xff1a;&＃xff08;i&＃xff09;监督学习&＃xff08;参数/非参数算法&＃xff0c;支持向量机&＃xff0c;内核&＃xff0c;神经网络&＃xff09;。&＃xff08;ii&＃xff09;无监督学习&＃xff08;聚类、降维、推荐系统、深度学习&＃xff09;。&＃xff08;iii&＃xff09;机器学习的最佳实践&＃xff08;偏差/方差理论;机器学习和人工智能的创新过程&＃xff09;。本课程还将借鉴众多案例研究和应用&＃xff0c;以便您还将学习如何应用学习算法来构建智能机器人&＃xff08;感知、控制&＃xff09;、文本理解&＃xff08;网络搜索、反垃圾邮件&＃xff09;、计算机视觉、医学信息学、音频、数据库挖掘和其他领域。

&＃x1f4da;2 运行结果

主函数部分代码&＃xff1a;

%% Machine Learning Online Class

%  Exercise 6 | Spam Classification with SVMs

%

%  Instructions

%  ------------

% 

%  This file contains code that helps you get started on the

%  exercise. You will need to complete the following functions:

%

%     gaussianKernel.m

%     dataset3Params.m

%     processEmail.m

%     emailFeatures.m

%

%  For this exercise, you will not need to change any code in this file,

%  or any other files other than those mentioned above.

%

%% Initialization

clear ; close all; clc

%% &＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61; Part 1: Email Preprocessing &＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;

%  To use an SVM to classify emails into Spam v.s. Non-Spam, you first need

%  to convert each email into a vector of features. In this part, you will

%  implement the preprocessing steps for each email. You should

%  complete the code in processEmail.m to produce a word indices vector

%  for a given email.

fprintf(&＃39;\nPreprocessing sample email (emailSample1.txt)\n&＃39;);

% Extract Features

file_contents &＃61; readFile(&＃39;emailSample1.txt&＃39;);

word_indices  &＃61; processEmail(file_contents);

% Print Stats

fprintf(&＃39;Word Indices: \n&＃39;);

fprintf(&＃39; %d&＃39;, word_indices);

fprintf(&＃39;\n\n&＃39;);

fprintf(&＃39;Program paused. Press enter to continue.\n&＃39;);

pause;

%% &＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61; Part 2: Feature Extraction &＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;

%  Now, you will convert each email into a vector of features in R^n. 

%  You should complete the code in emailFeatures.m to produce a feature

%  vector for a given email.

fprintf(&＃39;\nExtracting features from sample email (emailSample1.txt)\n&＃39;);

% Extract Features

file_contents &＃61; readFile(&＃39;emailSample1.txt&＃39;);

word_indices  &＃61; processEmail(file_contents);

features      &＃61; emailFeatures(word_indices);

% Print Stats

fprintf(&＃39;Length of feature vector: %d\n&＃39;, length(features));

fprintf(&＃39;Number of non-zero entries: %d\n&＃39;, sum(features > 0));

fprintf(&＃39;Program paused. Press enter to continue.\n&＃39;);

pause;

%% &＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61; Part 3: Train Linear SVM for Spam Classification &＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;

%  In this section, you will train a linear classifier to determine if an

%  email is Spam or Not-Spam.

% Load the Spam Email dataset

% You will have X, y in your environment

load(&＃39;spamTrain.mat&＃39;);

fprintf(&＃39;\nTraining Linear SVM (Spam Classification)\n&＃39;)

fprintf(&＃39;(this may take 1 to 2 minutes) ...\n&＃39;)

C &＃61; 0.1;

model &＃61; svmTrain(X, y, C, &＃64;linearKernel);

p &＃61; svmPredict(model, X);

fprintf(&＃39;Training Accuracy: %f\n&＃39;, mean(double(p &＃61;&＃61; y)) * 100);

%% &＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61; Part 4: Test Spam Classification &＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;

%  After training the classifier, we can evaluate it on a test set. We have

%  included a test set in spamTest.mat

% Load the test dataset

% You will have Xtest, ytest in your environment

load(&＃39;spamTest.mat&＃39;);

fprintf(&＃39;\nEvaluating the trained Linear SVM on a test set ...\n&＃39;)

p &＃61; svmPredict(model, Xtest);

fprintf(&＃39;Test Accuracy: %f\n&＃39;, mean(double(p &＃61;&＃61; ytest)) * 100);

pause;

%% &＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61; Part 5: Top Predictors of Spam &＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;

%  Since the model we are training is a linear SVM, we can inspect the

%  weights learned by the model to understand better how it is determining

%  whether an email is spam or not. The following code finds the words with

%  the highest weights in the classifier. Informally, the classifier

%  &＃39;thinks&＃39; that these words are the most likely indicators of spam.

%

% Sort the weights and obtin the vocabulary list

[weight, idx] &＃61; sort(model.w, &＃39;descend&＃39;);

vocabList &＃61; getVocabList();

fprintf(&＃39;\nTop predictors of spam: \n&＃39;);

for i &＃61; 1:15

    fprintf(&＃39; %-15s (%f) \n&＃39;, vocabList{idx(i)}, weight(i));

end

fprintf(&＃39;\n\n&＃39;);

fprintf(&＃39;\nProgram paused. Press enter to continue.\n&＃39;);

pause;

%% &＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61; Part 6: Try Your Own Emails &＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;&＃61;

%  Now that you&＃39;ve trained the spam classifier, you can use it on your own

%  emails! In the starter code, we have included spamSample1.txt,

%  spamSample2.txt, emailSample1.txt and emailSample2.txt as examples. 

%  The following code reads in one of these emails and then uses your 

%  learned SVM classifier to determine whether the email is Spam or 

%  Not Spam

% Set the file to be read in (change this to spamSample2.txt,

% emailSample1.txt or emailSample2.txt to see different predictions on

% different emails types). Try your own emails as well!

filename &＃61; &＃39;spamSample1.txt&＃39;;

% Read and predict

file_contents &＃61; readFile(filename);

word_indices  &＃61; processEmail(file_contents);

x             &＃61; emailFeatures(word_indices);

p &＃61; svmPredict(model, x);

fprintf(&＃39;\nProcessed %s\n\nSpam Classification: %d\n&＃39;, filename, p);

fprintf(&＃39;(1 indicates spam, 0 indicates not spam)\n\n&＃39;);

&＃x1f389;3 参考文献

​[1]谢宜鑫. 基于机器学习的建筑空调能耗数据挖掘和模式识别[D].北京交通大学,2019.

&＃x1f468;‍&＃x1f4bb;4 Matlab代码