Оглавление
John Paul Mueller. Machine Learning For Dummies
Machine Learning For Dummies® To view this book's Cheat Sheet, simply go to www.dummies.com and search for “Machine Learning For Dummies Cheat Sheet” in the Search box. Table of Contents
List of Tables
List of Illustrations
Guide
Pages
Introduction
About This Book
Foolish Assumptions
Icons Used in This Book
Beyond the Book
Where to Go from Here
Introducing How Machines Learn
Getting the Real Story about AI
Moving beyond the Hype
YES, FULLY AUTONOMOUS WEAPONS EXIST
Dreaming of Electric Sheep
Understanding the history of AI and machine learning
Exploring what machine learning can do for AI
Considering the goals of machine learning
Defining machine learning limits based on hardware
Overcoming AI Fantasies
Discovering the fad uses of AI and machine learning
Considering the true uses of AI and machine learning
Being useful; being mundane
Considering the Relationship between AI and Machine Learning
Considering AI and Machine Learning Specifications
Defining the Divide between Art and Engineering
Predicting the Next AI Winter
Learning in the Age of Big Data
Considering the Machine Learning Essentials
Defining Big Data
JUST HOW BIG IS BIG?
Considering the Sources of Big Data
Building a new data source
Obtaining data from public sources
Obtaining data from private sources
Creating new data from existing data
Using existing data sources
Locating test data sources
Specifying the Role of Statistics in Machine Learning
Understanding the Role of Algorithms
Defining what algorithms do
Considering the five main techniques
Symbolic reasoning
Connections modelled on the brain’s neurons
Evolutionary algorithms that test variation
Bayesian inference
Systems that learn by analogy
Defining What Training Means
Having a Glance at the Future
Creating Useful Technologies for the Future
Considering the role of machine learning in robots
Using machine learning in health care
Creating smart systems for various needs
Using machine learning in industrial settings
Understanding the role of updated processors and other hardware
Discovering the New Work Opportunities with Machine Learning
Working for a machine
Working with machines
Repairing machines
Creating new machine learning tasks
Devising new machine learning environments
Avoiding the Potential Pitfalls of Future Technologies
Preparing Your Learning Tools
Installing a Python Distribution
Using Anaconda for Machine Learning
Getting Anaconda
Defining why Anaconda is used in this book
Installing Anaconda on Linux
Installing Anaconda on Mac OS X
Installing Anaconda on Windows
A WORD ABOUT THE SCREENSHOTS
Downloading the Datasets and Example Code
Using Jupyter Notebook
Starting Jupyter Notebook
Stopping the Jupyter Notebook server
Defining the code repository
Defining the book’s folder
Creating a new notebook
Exporting a notebook
Removing a notebook
Importing a notebook
Understanding the datasets used in this book
Beyond Basic Coding in Python
Defining the Basics You Should Know
Considering Python basics
Working with functions
Creating reusable functions
Calling functions
SENDING REQUIRED ARGUMENTS
SENDING ARGUMENTS BY KEYWORD
GIVING FUNCTION ARGUMENTS A DEFAULT VALUE
CREATING FUNCTIONS WITH A VARIABLE NUMBER OF ARGUMENTS
Working with modules
Storing Data Using Sets, Lists, and Tuples
Creating sets
Performing operations on sets
Using lists
Defining a list
Combining lists using concatenation
Constructing lists using comprehensions
Slicing and dicing lists
Creating and using tuples
Defining Useful Iterators
Working with ranges
Iterating multiple lists using zip
Working with generators using yield
Indexing Data Using Dictionaries
Creating dictionaries
Storing and retrieving data from dictionaries
Working with Google Colab
Defining Google Colab
Understanding what Google Colab does
SOME FIREFOX ODDITIES
Considering the online coding difference
Using local runtime support
Working with Google Colab features
Getting a Google Account
Creating the account
Signing in
Working with Notebooks
Creating a new notebook
Opening existing notebooks
Using Google Drive for existing notebooks
Using GitHub for existing notebooks
Using local storage for existing notebooks
Uploading a notebook
Saving notebooks
Using Drive to save notebooks
Using GitHub to save notebooks
Using GitHub Gist to save notebooks
Downloading notebooks
Performing Common Tasks
Creating code cells
Creating text cells
Creating special cells
Working with headings
Working with a table of contents
Editing cells
Moving cells
Using Hardware Acceleration
Viewing Your Notebook
Displaying the table of contents
Getting notebook information
Checking code execution
Executing the Code
Sharing Your Notebook
Getting Help
Getting Started with the Math Basics
Demystifying the Math Behind Machine Learning
Working with Data
Learning the terminology
Understanding scalar and vector operations
Performing vector multiplication
Creating a matrix
Understanding basic operations
Performing matrix multiplication
Glancing at advanced matrix operations
Using vectorization effectively
Exploring the World of Probabilities
Getting an overview of probability
Operating on probabilities
Conditioning chance by Bayes' theorem
Describing the Use of Statistics
Descending the Gradient
Acknowledging Different Kinds of Learning
Supervised learning
Unsupervised learning
Reinforcement learning
The learning process
Mapping an unknown function
Exploring cost functions
Descending the optimization curve
Optimizing with big data
Leveraging sampling
Using parallelism
Learning out-of-core
Validating Machine Learning
Considering the Use of Example Data
Checking Out-of-Sample Errors
Understanding the concept of samples
Looking for the holy grail of generalization
Experimenting how bias and variance work
Keeping model complexity in mind
Keeping solutions balanced
Depicting learning curves
Training, Validating, and Testing
Considering the split
Resorting to cross-validation
Looking for alternatives in validation
Optimizing by Cross-Validation
Sources of predictive performance
Exploring the hyper-parameter space
Selecting relevant features
Avoiding Sample Bias and Leakage Traps
Starting with Simple Learners
Discovering the Incredible Perceptron
Falling short of a miracle
Hitting the nonseparability limit
Growing Greedy Classification Trees
Predicting outcomes by splitting data
Pruning overgrown trees
Taking a Probabilistic Turn
Understanding Naïve Bayes
Estimating response with Naïve Bayes
Learning from Smart and Big Data
Preprocessing Data
Gathering and Cleaning Data
Repairing Missing Data
Identifying missing data
Choosing the right replacement strategy
Transforming Distributions
Creating Your Own Features
Understanding the need to create features
Creating features automatically
Explaining the basics of SVD
Reorganizing data
Delimiting Anomalous Data
Using a univariate strategy
Resorting to Multivariate Models
Leveraging Similarity
Measuring Similarity between Vectors
Understanding similarity
Computing distances for learning
Euclidean distance
Manhattan distance
Chebyshev distance
Using Distances to Locate Clusters
Checking assumptions and expectations
Inspecting the gears of the K-means algorithm
Tuning the K-Means Algorithm
Experimenting with K-means reliability
Experimenting with how centroids converge
Finding Similarity by K-Nearest Neighbors
Understanding the k parameter
Experimenting with a flexible algorithm
Working with Linear Models the Easy Way
Starting to Combine Features
Getting an overview of regression
Solving problems with a machine learning approach
Understanding R squared
Mixing Features of Different Types
Switching to Probabilities
Specifying a binary response
Handling multiple classes
Guessing the Right Features
Defining the outcome of features that don’t work together
Solving overfitting by using greedy selection
Addressing overfitting by regularization
Learning One Example at a Time
Using gradient descent
Understanding how SGD is different
Hitting Complexity with Neural Networks
Revising the Perceptron
Pushing forth with feed-forward
Going even deeper down the rabbit hole
Pulling back with backpropagation
Representing the Way of Learning of a Network
Understanding the problem with overfitting
Choosing a framework
Getting your copy of TensorFlow and Keras
REDUCING CONDA AND PIP ERRORS
Opening the black box
Introducing Deep Learning
Understanding some deep learning essentials
Explaining the magic of convolutions
Understanding recurrent neural networks
Going a Step Beyond Using Support Vector Machines
Revisiting the Separation Problem
Explaining the Algorithm
Avoiding the pitfalls of nonseparability
Applying nonlinearity
Explaining the kernel trick by example
Classifying and Estimating with SVM
Resorting to Ensembles of Learners
Leveraging Decision Trees
Growing a forest of trees
Creating the Random Forests ensemble
GETTING MORE RF INFORMATION
Demonstrating the RF algorithm
Understanding the importance measures
Working with Almost Random Guesses
Bagging predictors with Adaboost
Boosting Smart Predictors
Meeting again with gradient descent
Considering the state of the art in tabular data
Averaging Different Predictors
Blending solutions
Stacking diverse solutions
Applying Learning to Real Problems
Classifying Images
Working with a Set of Images
Revising the State of the Art in Computer Vision
Extracting Visual Features
Recognizing Faces Using Eigenfaces
Classifying Images
Scoring Opinions and Sentiments
Introducing Natural Language Processing
Revising the State of the Art in NLP
Understanding How Machines Read
Defining the input data
Processing and enhancing text
Considering basic processing tasks
Stemming and removing stop words
Scraping textual datasets from the web
Handling problems with raw text
Using Scoring and Classification
Performing classification tasks
Analyzing reviews from e-commerce
Recommending Products and Movies
Realizing the Revolution of E-Commerce
Downloading Rating Data
Trudging through the MovieLens dataset
Navigating through anonymous web data
Encountering the limits of rating data
Considering collaborative filtering
Massaging the data
Performing collaborative filtering
Catching the Limits of Behavioral Data
Integrating Text and Behaviors
Viewing the attributes
Obtaining statistics
Leveraging SVD
Understanding the SVD connection
Seeing SVD in action
The Part of Tens
Ten Ways to Improve Your Machine Learning Models
Studying Learning Curves
Using Cross-Validation Correctly
Choosing the Right Error or Score Metric
Searching for the Best Hyper-Parameters
Testing Multiple Models
Averaging Models
Stacking Models
Applying Feature Engineering
Selecting Features and Examples
Looking for More Data
Ten Guidelines for Ethical Data Usage
Obtaining Permission
Using Sanitization Techniques
Avoiding Data Inference
Using Generalizations Correctly
Shunning Discriminatory Practices
Detecting Black Swans in Code
Understanding the Process
Considering the Consequences of an Action
Balancing Decision Making
Verifying a Data Source
Ten Machine Learning Packages to Master
Gensim
imbalanced-learn
OpenCV
SciPy
SHAP
Statsmodels
Modin
PyTorch
Poetry
Snorkel
Index. A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Z
About the Authors
John’s Dedication
Luca’s Dedication
John’s Acknowledgments
Luca’s Acknowledgments
WILEY END USER LICENSE AGREEMENT
