Python : deeper insights into machine learning : leverage benefits of machine learning techniques using Python : a course in three modules

Leverage benefits of machine learning techniques using Python About This Book Improve and optimise machine learning systems using effective strategies. Develop a strategy to deal with a large amount of data. Use of Python code for implementing a range of machine learning algorithms and techniques. Who This Book Is For This title is for data scientist and researchers who are already into the field of data science and want to see machine learning in action and explore its real-world application. Prior knowledge of Python programming and mathematics is must with basic knowledge of machine learning concepts. What You Will Learn Learn to write clean and elegant Python code that will optimize the strength of your algorithms Uncover hidden patterns and structures in data with clustering Improve accuracy and consistency of results using powerful feature engineering techniques Gain practical and theoretical understanding of cutting-edge deep learning algorithms Solve unique tasks by building models Get grips on the machine learning design process In Detail Machine learning and predictive analytics are becoming one of the key strategies for unlocking growth in a challenging contemporary marketplace. It is one of the fastest growing trends in modern computing, and everyone wants to get into the field of machine learning. In order to obtain sufficient recognition in this field, one must be able to understand and design a machine learning system that serves the needs of a project. The idea is to prepare a learning path that will help you to tackle the real-world complexities of modern machine learning with innovative and cutting-edge techniques. Also, it will give you a solid foundation in the machine learning design process, and enable you to build customized machine learning models to solve unique problems. The course begins with getting your Python fundamentals nailed down. It focuses on answering the right questions that cove a wide range of powerful Python libraries, including scikit-learn Theano and Keras.After getting familiar with Python core concepts, it's time to dive into the field of data science. You will further gain a solid foundation on the machine learning design and also learn to customize models for solving problems. At a later stage, you will get a grip on more advanced techniques and acquire a broad set of powerful skills in the area of feature selection and feature engineering. Style and approach This course includes all the resources that will help you jump into the data science field with Python. The aim is to walk through the elements of Python covering powerful machine learning libraries. This course will explain important machine learning models in a step-by-step manner. Each topic is well explained with real-world applications with detailed guidance.Through this comprehensive guide, you will be able to explore machine learning techniques. Cover......Page 1 Copyright......Page 3 Credits......Page 4 Preface......Page 6 Table of Contents......Page 10 Giving Computers the Ability to Learn from Data......Page 20 The three different types of machine learning......Page 21 An introduction to the basic terminology and notations......Page 27 A roadmap for building machine learning systems......Page 29 Using Python for machine learning......Page 32 Summary......Page 34 Training Machine Learning Algorithms for Classification......Page 36 Artificial neurons – a brief glimpse into the early history of machine learning......Page 37 Implementing a perceptron learning algorithm in Python......Page 43 Adaptive linear neurons and the convergence of learning......Page 52 Summary......Page 66 Choosing a classification algorithm......Page 68 First steps with scikit-learn......Page 69 Modeling class probabilities via logistic regression......Page 75 Maximum margin classification with support vector machines......Page 88 Solving nonlinear problems using a kernel SVM......Page 94 Decision tree learning......Page 99 K-nearest neighbors – a lazy learning algorithm......Page 111 Summary......Page 115 Dealing with missing data......Page 118 Handling categorical data......Page 123 Partitioning a dataset in training and test sets......Page 127 Bringing features onto the same scale......Page 129 Selecting meaningful features......Page 131 Assessing feature importance with random forests......Page 143 Summary......Page 145 Compressing Data via Dimensionality Reduction......Page 146 Unsupervised dimensionality reduction via principal component analysis......Page 147 Supervised data compression via linear discriminant analysis......Page 157 Using kernel principal component analysis for nonlinear mappings......Page 167 Summary......Page 186 Streamlining workflows with pipelines......Page 188 Using k-fold cross-validation to assess model performance......Page 192 Debugging algorithms with learning and validation curves......Page 198 Fine-tuning machine learning models via grid search......Page 204 Looking at different performance evaluation metrics......Page 208 Summary......Page 217 Learning with ensembles......Page 218 Implementing a simple majority vote classifier......Page 222 Evaluating and tuning the ensemble classifier......Page 232 Bagging – building an ensemble of classifiers from bootstrap samples......Page 238 Leveraging weak learners via adaptive boosting......Page 243 Summary......Page 251 Obtaining the IMDb movie review dataset......Page 252 Introducing the bag-of-words model......Page 255 Training a logistic regression model for document classification......Page 263 Working with bigger data – online algorithms and out-of-core learning......Page 265 Summary......Page 269 Embedding a Machine Learning Model into a Web Application......Page 270 Serializing fitted scikit-learn estimators......Page 271 Setting up a SQLite database for data storage......Page 274 Developing a web application with Flask......Page 276 Turning the movie classifier into a web application......Page 283 Deploying the web application to a public server......Page 291 Summary......Page 295 Predicting Continuous Target Variables with Regression Analysis......Page 296 Introducing a simple linear regression model......Page 297 Exploring the Housing Dataset......Page 298 Implementing an ordinary least squares linear regression model......Page 304 Fitting a robust regression model using RANSAC......Page 310 Evaluating the performance of linear regression models......Page 313 Using regularized methods for regression......Page 316 Turning a linear regression model into a curve – polynomial regression......Page 317 Summary......Page 328 Working with Unlabeled Data – Clustering Analysis......Page 330 Grouping objects by similarity using k-means......Page 331 Organizing clusters as a hierarchical tree......Page 345 Locating regions of high density via DBSCAN......Page 353 Summary......Page 359 Training Artificial Neural Networks for Image Recognition......Page 360 Modeling complex functions with artificial neural networks......Page 361 Classifying handwritten digits......Page 369 Training an artificial neural network......Page 384 Developing your intuition for backpropagation......Page 391 Debugging neural networks with gradient checking......Page 392 Convergence in neural networks......Page 398 Other neural network architectures......Page 400 A few last words about neural network implementation......Page 403 Summary......Page 404 Parallelizing Neural Network Training with Theano......Page 406 Building, compiling, and running expressions with Theano......Page 407 Choosing activation functions for feedforward neural networks......Page 420 Training neural networks efficiently using Keras......Page 427 Summary......Page 433 Thinking in Machine Learning......Page 438 The human interface......Page 439 Design principles......Page 442 Summary......Page 470 Tools and Techniques......Page 472 IPython console......Page 473 Installing the SciPy stack......Page 474 NumPY......Page 475 Matplotlib......Page 481 Pandas......Page 485 SciPy......Page 488 Scikit-learn......Page 491 Summary......Page 498 Turning Data into Information......Page 500 Big data......Page 501 Signals......Page 517 Cleaning data......Page 519 Visualizing data......Page 521 Summary......Page 524 Logical models......Page 526 Tree models......Page 534 Rule models......Page 538 Summary......Page 545 Linear Models......Page 546 Introducing least squares......Page 547 Logistic regression......Page 555 Multiclass classification......Page 561 Regularization......Page 562 Summary......Page 565 Getting started with neural networks......Page 566 Logistic units......Page 568 Cost function......Page 573 Implementing a neural network......Page 576 Gradient checking......Page 582 Other neural net architectures......Page 583 Summary......Page 584 Features – How Algorithms See the World......Page 586 Feature types......Page 587 Operations and statistics......Page 588 Transforming features......Page 591 Principle component analysis......Page 600 Summary......Page 602 Ensemble types......Page 604 Bagging......Page 605 Boosting......Page 611 Ensemble strategies......Page 618 Summary......Page 621 Evaluating model performance......Page 622 Model selection......Page 627 Learning curves......Page 630 Real-world case studies......Page 632 Machine learning at a glance......Page 643 Summary......Page 644 Unsupervised Machine Learning......Page 648 Principal component analysis......Page 649 Introducing k-means clustering......Page 654 Self-organizing maps......Page 665 Further reading......Page 671 Summary......Page 672 Deep Belief Networks......Page 674 Neural networks – a primer......Page 675 Restricted Boltzmann Machine......Page 680 Deep belief networks......Page 696 Further reading......Page 702 Summary......Page 703 Autoencoders......Page 704 Stacked Denoising Autoencoders......Page 713 Summary......Page 722 Introducing the CNN......Page 724 Further Reading......Page 746 Summary......Page 747 Introduction......Page 748 Understanding semi-supervised learning......Page 749 Semi-supervised algorithms in action......Page 750 Further reading......Page 773 Summary......Page 774 Introduction......Page 776 Text feature engineering......Page 777 Further reading......Page 800 Summary......Page 801 Introduction......Page 802 Creating a feature set......Page 803 Feature engineering in practice......Page 822 Further reading......Page 846 Summary......Page 847 Ensemble Methods......Page 848 Introducing ensembles......Page 849 Using models in dynamic applications......Page 868 Further reading......Page 880 Summary......Page 881 Additional Python Machine Learning Tools......Page 882 Alternative development tools......Page 883 Summary......Page 892 Chapter Code Requirements......Page 896 Biblography......Page 898 Cover 1 Copyright 3 Credits 4 Preface 6 Table of Contents 10 Giving Computers the Ability to Learn from Data 20 Building intelligent machines to transform data into knowledge 21 The three different types of machine learning 21 An introduction to the basic terminology and notations 27 A roadmap for building machine learning systems 29 Using Python for machine learning 32 Summary 34 Training Machine Learning Algorithms for Classification 36 Artificial neurons – a brief glimpse into the early history of machine learning 37 Implementing a perceptron learning algorithm in Python 43 Adaptive linear neurons and the convergence of learning 52 Summary 66 A Tour of Machine Learning Classifiers Using Scikit-learn 68 Choosing a classification algorithm 68 First steps with scikit-learn 69 Modeling class probabilities via logistic regression 75 Maximum margin classification with support vector machines 88 Solving nonlinear problems using a kernel SVM 94 Decision tree learning 99 K-nearest neighbors – a lazy learning algorithm 111 Summary 115 Building Good Training Sets – Data Preprocessing 118 Dealing with missing data 118 Handling categorical data 123 Partitioning a dataset in training and test sets 127 Bringing features onto the same scale 129 Selecting meaningful features 131 Assessing feature importance with random forests 143 Summary 145 Compressing Data via Dimensionality Reduction 146 Unsupervised dimensionality reduction via principal component analysis 147 Supervised data compression via linear discriminant analysis 157 Using kernel principal component analysis for nonlinear mappings 167 Summary 186 Learning Best Practices for Model Evaluation and Hyperparameter Tuning 188 Streamlining workflows with pipelines 188 Using k-fold cross-validation to assess model performance 192 Debugging algorithms with learning and validation curves 198 Fine-tuning machine learning models via grid search 204 Looking at different performance evaluation metrics 208 Summary 217 Combining Different Models for Ensemble Learning 218 Learning with ensembles 218 Implementing a simple majority vote classifier 222 Evaluating and tuning the ensemble classifier 232 Bagging – building an ensemble of classifiers from bootstrap samples 238 Leveraging weak learners via adaptive boosting 243 Summary 251 Applying Machine Learning to Sentiment Analysis 252 Obtaining the IMDb movie review dataset 252 Introducing the bag-of-words model 255 Training a logistic regression model for document classification 263 Working with bigger data – online algorithms and out-of-core learning 265 Summary 269 Embedding a Machine Learning Model into a Web Application 270 Serializing fitted scikit-learn estimators 271 Setting up a SQLite database for data storage 274 Developing a web application with Flask 276 Turning the movie classifier into a web application 283 Deploying the web application to a public server 291 Summary 295 Predicting Continuous Target Variables with Regression Analysis 296 Introducing a simple linear regression model 297 Exploring the Housing Dataset 298 Implementing an ordinary least squares linear regression model 304 Fitting a robust regression model using RANSAC 310 Evaluating the performance of linear regression models 313 Using regularized methods for regression 316 Turning a linear regression model into a curve – polynomial regression 317 Summary 328 Working with Unlabeled Data – Clustering Analysis 330 Grouping objects by similarity using k-means 331 Organizing clusters as a hierarchical tree 345 Locating regions of high density via DBSCAN 353 Summary 359 Training Artificial Neural Networks for Image Recognition 360 Modeling complex functions with artificial neural networks 361 Classifying handwritten digits 369 Training an artificial neural network 384 Developing your intuition for backpropagation 391 Debugging neural networks with gradient checking 392 Convergence in neural networks 398 Other neural network architectures 400 A few last words about neural network implementation 403 Summary 404 Parallelizing Neural Network Training with Theano 406 Building, compiling, and running expressions with Theano 407 Choosing activation functions for feedforward neural networks 420 Training neural networks efficiently using Keras 427 Summary 433 Thinking in Machine Learning 438 The human interface 439 Design principles 442 Summary 470 Tools and Techniques 472 Python for machine learning 473 IPython console 473 Installing the SciPy stack 474 NumPY 475 Matplotlib 481 Pandas 485 SciPy 488 Scikit-learn 491 Summary 498 Turning Data into Information 500 What is data? 501 Big data 501 Signals 517 Cleaning data 519 Visualizing data 521 Summary 524 Models – Learning from Information 526 Logical models 526 Tree models 534 Rule models 538 Summary 545 Linear Models 546 Introducing least squares 547 Logistic regression 555 Multiclass classification 561 Regularization 562 Summary 565 Neural Networks 566 Getting started with neural networks 566 Logistic units 568 Cost function 573 Implementing a neural network 576 Gradient checking 582 Other neural net architectures 583 Summary 584 Features – How Algorithms See the World 586 Feature types 587 Operations and statistics 588 Structured features 591 Transforming features 591 Principle component analysis 600 Summary 602 Learning with Ensembles 604 Ensemble types 604 Bagging 605 Boosting 611 Ensemble strategies 618 Summary 621 Design Strategies and Case Studies 622 Evaluating model performance 622 Model selection 627 Learning curves 630 Real-world case studies 632 Machine learning at a glance 643 Summary 644 Unsupervised Machine Learning 648 Principal component analysis 649 Introducing k-means clustering 654 Self-organizing maps 665 Further reading 671 Summary 672 Deep Belief Networks 674 Neural networks – a primer 675 Restricted Boltzmann Machine 680 Deep belief networks 696 Further reading 702 Summary 703 Stacked Denoising Autoencoders 704 Autoencoders 704 Stacked Denoising Autoencoders 713 Further reading 722 Summary 722 Convolutional Neural Networks 724 Introducing the CNN 724 Further Reading 746 Summary 747 Semi-Supervised Learning 748 Introduction 748 Understanding semi-supervised learning 749 Semi-supervised algorithms in action 750 Further reading 773 Summary 774 Text Feature Engineering 776 Introduction 776 Text feature engineering 777 Further reading 800 Summary 801 Feature Engineering Part II 802 Introduction 802 Creating a feature set 803 Feature engineering in practice 822 Further reading 846 Summary 847 Ensemble Methods 848 Introducing ensembles 849 Using models in dynamic applications 868 Further reading 880 Summary 881 Additional Python Machine Learning Tools 882 Alternative development tools 883 Further reading 892 Summary 892 Chapter Code Requirements 896 Biblography 898 Leverage benefits of machine learning techniques using Python About This Book • Improve and optimise machine learning systems using effective strategies. • Develop a strategy to deal with a large amount of data. • Use of Python code for implementing a range of machine learning algorithms and techniques. Who This Book Is For This title is for data scientist and researchers who are already into the field of data science and want to see machine learning in action and explore its real-world application. Prior knowledge of Python programming and mathematics is must with basic knowledge of machine learning concepts. What You Will Learn • Learn to write clean and elegant Python code that will optimize the strength of your algorithms • Uncover hidden patterns and structures in data with clustering • Improve accuracy and consistency of results using powerful feature engineering techniques • Gain practical and theoretical understanding of cutting-edge deep learning algorithms • Solve unique tasks by building models • Get grips on the machine learning design process In Detail Machine learning and predictive analytics are becoming one of the key strategies for unlocking growth in a challenging contemporary marketplace. It is one of the fastest growing trends in modern computing, and everyone wants to get into the field of machine learning. In order to obtain sufficient recognition in this field, one must be able to understand and design a machine learning system that serves the needs of a project. The idea is to prepare a learning path that will help you to tackle the real-world complexities of modern machine learning with innovative and cutting-edge techniques. Also, it will give you a solid foundation in the machine learning design process, and enable you to build customized machine learning models to solve unique problems. The course begins with getting your Python fundamentals nailed down. It focuses on answering the right questions that cove a wide range of powerful Python libraries, including scikit-learn Theano and Keras.After getting familiar with Python core concepts, it's time to dive into the field of data science. You will further gain a solid foundation on the machine learning design and also learn to customize models for solving problems. At a later stage, you will get a grip on more advanced techniques and acquire a broad set of powerful skills in the area of feature selection and feature engineering. Style and approach This course includes all the resources that will help you jump into the data science field with Python. The aim is to walk through the elements of Python covering powerful machine learning libraries. This course will explain important machine learning models in a step-by-step manner. Each topic is well explained with real-world applications with detailed guidance.Through this comprehensive guide, you will be able to explore machine learning techniques. Leverage benefits of machine learning techniques using PythonAbout This Book Improve and optimise machine learning systems using effective strategies. Develop a strategy to deal with a large amount of data. Use of Python code for implementing a range of machine learning algorithms and techniques. Who This Book Is For This title is for data scientist and researchers who are already into the field of data science and want to see machine learning in action and explore its real-world application. Prior knowledge of Python programming and mathematics is must with basic knowledge of machine learning concepts. What You Will Learn Learn to write clean and elegant Python code that will optimize the strength of your algorithms Uncover hidden patterns and structures in data with clustering Improve accuracy and consistency of results using powerful feature engineering techniques Gain practical and theoretical understanding of cutting-edge deep learning algorithms Solve unique tasks by building models Get grips on the machine learning design processIn Detail Machine learning and predictive analytics are becoming one of the key strategies for unlocking growth in a challenging contemporary marketplace. It is one of the fastest growing trends in modern computing, and everyone wants to get into the field of machine learning. In order to obtain sufficient recognition in this field, one must be able to understand and design a machine learning system that serves the needs of a project. The idea is to prepare a learning path that will help you to tackle the real-world complexities of modern machine learning with innovative and cutting-edge techniques. Also, it will give you a solid foundation in the machine learning design process, and enable you to build customized machine learning models to solve unique problems. The course begins with getting your Python fundamentals nailed down. It focuses on answering the right questions that cove a wide range of powerful Python libraries, including scikit-learn Theano and Keras. After getting familiar with Python core concepts, it's time to dive into the field of data science. You will further gain a solid foundation on the machine learning design and also learn to customize models for solving problems. At a later stage, you will get a grip on more advanced techniques and acquire a broad set of powerful skills in the area of feature selection and feature engineering. Style and approach This course includes all the resources that will help you jump into the data science field with Python. The aim is to walk through the elements of Python covering powerful machine learning libraries. This course will explain important machine learning models in a step-by-step manner. Each topic is well explained with real-world applications with detailed guidance. Through this comprehensive guide, you will be able to explore machine learning techniques Leverage benefits of machine learning techniques using Python. About This Book Improve and optimise machine learning systems using effective strategies. Develop a strategy to deal with a large amount of data. Use of Python code for implementing a range of machine learning algorithms and techniques. Who This Book Is For This title is for data scientist and researchers who are already into the field of data science and want to see machine learning in action and explore its real-world application. Prior knowledge of Python programming and mathematics is must with basic knowledge of machine learning concepts. What You Will Learn Learn to write clean and elegant Python code that will optimize the strength of your algorithms Uncover hidden patterns and structures in data with clustering Improve accuracy and consistency of results using powerful feature engineering techniques Gain practical and theoretical understanding of cutting-edge deep learning algorithms Solve unique tasks by building models Get grips on the machine learning design process In Detail Machine learning and predictive analytics are becoming one of the key strategies for unlocking growth in a challenging contemporary marketplace. It is one of the fastest growing trends in modern computing, and everyone wants to get into the field of machine learning. In order to obtain sufficient recognition in this field, one must be able to understand and design a machine learning system that serves the needs of a project. The idea is to prepare a learning path that will help you to tackle the real-world complexities of modern machine learning with innovative and cutting-edge techniques. Also, it will give you a solid foundation in the machine learning design process, and enable you to build customized machine learning models to solve unique problems. The course begins with getting your Python fundamentals nailed down. It focuses on answering the right questions that cove a wide range of powerful Python libraries, including scikit-learn Theano and Keras. After getting familiar with Python core concepts, it's time to dive into the field of data science. You will further gain a solid foundation on the machine learning design and also learn to customize models for solving problems. At a later stage, you will get a grip on more advanced techniques and acquire a broad set of powerful skills in the area of feature selection and feature engineering. Style and approach This course includes all the resource .. Edited By John S. Najarian, John P. Delaney. Includes Bibliographical References And Index.