book

Grokking Machine Learning

by Luis Serrano

December 2021

Intermediate to advanced

512 pages

15h 23m

English

Manning Publications

Read now

Unlock full access

forewordprefaceacknowledgmentsabout this bookTypes of chaptersRecommended learning pathsAppendicesRequirements and learning goalsOther resourcesWe’ll be writing codeabout the author
Do I need a heavy math and coding background to understand machine learning?OK, so what exactly is machine learning?How do we get machines to make decisions with data? The remember-formulate-predict frameworkSummary
What is the difference between labeled and unlabeled data?Supervised learning: The branch of machine learning that works with labeled dataUnsupervised learning: The branch of machine learning that works with unlabeled dataWhat is reinforcement learning?SummaryExercises
The problem: We need to predict the price of a houseThe solution: Building a regression model for housing pricesHow to get the computer to draw this line: The linear regression algorithmHow do we measure our results? The error functionReal-life application: Using Turi Create to predict housing prices in IndiaWhat if the data is not in a line? Polynomial regressionParameters and hyperparametersApplications of regressionSummaryExercises
An example of underfitting and overfitting using polynomial regressionHow do we get the computer to pick the right model? By testingWhere did we break the golden rule, and how do we fix it? The validation setA numerical way to decide how complex our model should be: The model complexity graphAnother alternative to avoiding overfitting: RegularizationPolynomial regression, testing, and regularization with Turi CreateSummaryExercises
The problem: We are on an alien planet, and we don’t know their language!How do we determine whether a classifier is good or bad? The error functionHow to find a good classifier? The perceptron algorithmCoding the perceptron algorithmApplications of the perceptron algorithmSummaryExercises

Logistic classifiers: A continuous version of perceptron classifiersHow to find a good logistic classifier? The logistic regression algorithmCoding the logistic regression algorithmReal-life application: Classifying IMDB reviews with Turi CreateClassifying into multiple classes: The softmax functionSummaryExercises
Accuracy: How often is my model correct?How to fix the accuracy problem? Defining different types of errors and how to measure themA useful tool to evaluate our model: The receiver operating characteristic (ROC) curveSummaryExercises
Sick or healthy? A story with Bayes’ theorem as the heroUse case: Spam-detection modelBuilding a spam-detection model with real dataSummaryExercises
The problem: We need to recommend apps to users according to what they are likely to downloadThe solution: Building an app-recommendation systemBeyond questions like yes/noThe graphical boundary of decision treesReal-life application: Modeling student admissions with Scikit-LearnDecision trees for regressionApplicationsSummaryExercises
Neural networks with an example: A more complicated alien planetTraining neural networksCoding neural networks in KerasNeural networks for regressionOther architectures for more complex datasetsSummaryExercises
Using a new error function to build better classifiersCoding support vector machines in Scikit-LearnTraining SVMs with nonlinear boundaries: The kernel methodSummaryExercises
With a little help from our friendsBagging: Joining some weak learners randomly to build a strong learnerAdaBoost: Joining weak learners in a clever way to build a strong learnerGradient boosting: Using decision trees to build strong learnersXGBoost: An extreme way to do gradient boostingApplications of ensemble methodsSummaryExercises
The Titanic datasetCleaning up our dataset: Missing values and how to deal with themFeature engineering: Transforming the features in our dataset before training the modelsTraining our modelsTuning the hyperparameters to find the best model: Grid searchUsing K-fold cross-validation to reuse our data as training and validationSummaryExercises
Chapter 2: Types of machine learningChapter 3: Drawing a line close to our points: Linear regressionChapter 4: Optimizing the training process: Underfitting, overfitting, testing, and regularizationChapter 5: Using lines to split our points: The perceptron algorithmChapter 6: A continuous approach to splitting points: Logistic classifiersChapter 7: How do you measure classification models? Accuracy and its friendsChapter 8: Using probability to its maximum: The naive Bayes modelChapter 9: Splitting data by asking questions: Decision treesChapter 10: Combining building blocks to gain more power: Neural networksChapter 11: Finding boundaries with style: Support vector machines and the kernel methodChapter 12: Combining models to maximize results: Ensemble learningChapter 13: Putting it all in practice: A real-life example of data engineering and machine learning
Using gradient descent to decrease functionsUsing gradient descent to train modelsUsing gradient descent for regularizationGetting stuck on local minima: How it happens, and how we solve it
General referencesCoursesBlogs and YouTube channelsBooksChapter 1Chapter 2Chapter 3Chapter 4Chapter 5Chapter 6Chapter 7Chapter 8Chapter 9Chapter 10Chapter 11Chapter 12Chapter 13Graphics and image icons

Content preview from Grokking Machine Learning

7 How do you measure classification models? Accuracy and its friends

In this chapter

types of errors a model can make: false positives and false negatives
putting these errors in a table: the confusion matrix
what are accuracy, recall, precision, F-score, sensitivity, and specificity, and how are they used to evaluate models
what is the ROC curve, and how does it keep track of sensitivity and specificity at the same time

This chapter is slightly different from the previous two—it doesn’t focus on building classification models; instead, it focuses on evaluating them. For a machine learning professional, being able to evaluate the performance ...