Index

A

• access keys
  • about / Downloading data from a social network
• accuracy
  • improving, dictionary used / Improving accuracy using a dictionary
• activation function
  • about / Artificial neural networks
• Adult dataset
  • URL / Representing reality in models
• Advertisements dataset
  • URL / Feature creation
• affinity analysis
  • example / A simple affinity analysis example
  • defining / What is affinity analysis?
  • product recommendations / Product recommendations
  • dataset, loading with NumPy / Loading the dataset with NumPy
  • ranking of rules, implementing / Implementing a simple ranking of rules
  • ranking, to find best rules / Ranking to find the best rules
  • about / Affinity analysis
  • algorithms / Algorithms for affinity analysis
  • parameters, selecting / Choosing parameters
• Amazon S3 console
  • URL / Training on Amazon's EMR infrastructure
• API endpoint
  • URL / Using a Web API to get data
• application
  • defining / Application
  • word counts, extracting / Extracting word counts
  • dictionaries, converting to matrix / Converting dictionaries to a matrix
  • Naive Bayes classifier, training / Training the Naive Bayes classifier
  • about / Application, Application
  • data, obtaining / Getting the data, Getting the data
  • neural network, creating / Creating the neural network
  • neural network, training with training dataset / Putting it all together
  • Naive Bayes algorithm / Naive Bayes prediction
• apps, Twitter account
  • URL / Downloading data from a social network
• Apriori algorithm / The Apriori algorithm
• Apriori implementation
  • about / The Apriori implementation
  • Apriori algorithm / The Apriori algorithm
  • defining / Implementation
• arbitrary websites
  • text, extracting from / Extracting text from arbitrary websites, Putting it all together
  • stories, finding / Finding the stories in arbitrary websites
  • data mining, using / Putting it all together
  • nodes, ignoring / Putting it all together
  • HTML file, parsing / Putting it all together
• Artificial Neural Networks
  • about / Artificial neural networks
• association rules
  • extracting / Extracting association rules
  • evaluating / Evaluation
• authorship analysis
  • defining / Attributing documents to authors
  • applications / Applications and use cases
  • use cases / Applications and use cases
  • about / Applications and use cases
  • authorship attribution / Attributing authorship
  • data, obtaining / Getting the data
• authorship analysis, problems
  • authorship profiling / Attributing documents to authors
  • authorship verification / Attributing documents to authors
  • authorship clustering / Attributing documents to authors
• authorship attribution / Attributing authorship
• AWS CLI
  • installing / Training on Amazon's EMR infrastructure
• AWS console
  • URL / Running our code on a GPU

B

• back propagation (backprop) algorithm / Back propagation
• bagging
  • about / Random forests
• BatchIterator instance
  • creating / Creating the neural network
• Bayes' theorem / Bayes' theorem
  • about / Bayes' theorem
  • equation / Bayes' theorem
• bias
  • about / How do ensembles work?
• big data
  • about / Big data
  • use cases / Application scenario and goals
• Bleeding Edge code
  • installing / Scalability with the nearest neighbor
  • URL / Scalability with the nearest neighbor
• blog posts
  • extracting / Extracting the blog posts
• blogs dataset
  • about / Blogs dataset

C

• CAPTCHA
  • creating / Drawing basic CAPTCHAs
• CAPTCHAs
  • references / Better (worse?) CAPTCHAs
  • defining / Better (worse?) CAPTCHAs
• CART (Classification and Regression Trees)
  • about / Decision trees
• character n-grams
  • about / Character n-grams
  • extracting / Extracting character n-grams
• CIFAR-10
  • about / Application scenario and goals
  • URL / Application scenario and goals
• class
  • about / A simple classification example
• classification
  • example / A simple classification example
  • about / What is classification?
  • examples / What is classification?
  • dataset, loading / Loading and preparing the dataset
  • dataset, preparing / Loading and preparing the dataset
  • OneR algorithm, implementing / Implementing the OneR algorithm
  • algorithm, testing / Testing the algorithm
• classifiers
  • comparing / Comparing classifiers
• closed problem
  • about / Attributing authorship
• cluster evaluation
  • URL / Evaluating the results
• clustering
  • about / Grouping news articles
• coassociation matrix
  • defining / Evidence accumulation
• complex algorithms
  • references / More complex algorithms
• complex features
  • references / More complex features
• confidence
  • about / Implementing a simple ranking of rules
  • computing / Implementing a simple ranking of rules
• connected components
  • about / Connected components
• Cosine distance
  • about / Distance metrics
• Coursera
  • about / More resources
  • references / More resources
• Coval font I, Open Font Library
  • URL / Drawing basic CAPTCHAs
• CPU
  • defining / When to use GPUs for computation
• cross-fold validation framework
  • defining / Running the algorithm
• CSV (Comma Separated Values)
  • about / Collecting the data

D

• data, blogging
  • URL / Getting the data
• data, Corpus
  • URL / Getting the data
• Dataframe
  • about / Using pandas to load the dataset
• data mining
  • defining / Introducing data mining
• dataset
  • loading / Loading the dataset, Loading the dataset, An introduction to Lasagne
  • data, collecting / Collecting the data
  • URL / Collecting the data
  • loading, pandas used / Using pandas to load the dataset
  • cleaning up / Cleaning up the dataset
  • new features, extracting / Extracting new features
  • classifying, with existing model / Classifying with an existing model
  • follower information, obtaining from Twitter / Getting follower information from Twitter
  • network, building / Building the network
  • graph, creating / Creating a graph
  • Similarity graph, creating / Creating a similarity graph
  • creating / Creating the dataset
  • CAPTCHAs, drawing / Drawing basic CAPTCHAs
  • image, splitting into individual letters / Splitting the image into individual letters
  • training dataset, creating / Creating a training dataset
  • training dataset, adjusting to methodology / Adjusting our training dataset to our methodology
• datasets
  • about / Introducing data mining
  • samples / Introducing data mining
  • features / Introducing data mining
  • example / Introducing data mining
  • URL / Obtaining the dataset, Extending the IPython Notebook
  • references / New datasets
• decision tree implementation
  • min_samples_split / Parameters in decision trees
  • min_samples_leaf / Parameters in decision trees
• decision trees
  • about / Decision trees
  • parameters / Parameters in decision trees
  • Gini impurity / Parameters in decision trees
  • Information gain / Parameters in decision trees
  • using / Using decision trees
• dictionary
  • used, for improving accuracy / Improving accuracy using a dictionary
  • ranking mechanisms, for words / Ranking mechanisms for words
  • improved prediction function, testing / Putting it all together
• DictVectorizer class
  • about / Converting dictionaries to a matrix
• disambiguation
  • about / Disambiguation
  • data, downloading from social network / Downloading data from a social network
  • dataset, loading / Loading and classifying the dataset
  • dataset, classifying / Loading and classifying the dataset
  • replicable dataset, creating from Twitter / Creating a replicable dataset from Twitter
• discretization
  • about / Common feature patterns
• discretization algorithm
  • defining / Loading and preparing the dataset
• documents
  • attributing, to authors / Attributing documents to authors

E

• EC2 service console
  • URL / Running our code on a GPU
• Eclat algorithm
  • about / Algorithms for affinity analysis
  • URL / The Eclat algorithm
  • implementing / The Eclat algorithm
• Elastic Map Reduce (EMR)
  • about / Training on Amazon's EMR infrastructure
• Enron dataset
  • using / Using the Enron dataset
  • accessing / Accessing the Enron dataset
  • URL / Accessing the Enron dataset
  • dataset loader, creating / Creating a dataset loader
  • existing parameter space, using / Putting it all together
  • classifier, using / Putting it all together
  • evaluation / Evaluation
• ensembles
  • clustering / Clustering ensembles
  • evidence accumulation / Evidence accumulation
  • working / How it works
  • implementing / Implementation
• environment
  • setting up / Setting up the environment
• epochs
  • about / Back propagation
• Euclidean distance
  • about / Distance metrics
• evaluation, of clustering algorithms
  • references / Evaluation
• Evidence Accumulation Clustering (EAC)
  • about / Evidence accumulation
  • defining / Evidence accumulation
• Excel, pandas
  • URL / More on pandas

F

• f1-score
  • about / Evaluation using the F1-score
  • computing / Evaluation using the F1-score
  • using / Evaluation using the F1-score
• feature-based normalization
  • about / Standard preprocessing
• feature creation
  • about / Feature creation
  • Principal Component Analysis (PCA) / Principal Component Analysis
• feature extraction
  • about / Feature extraction
  • reality, representing in models / Representing reality in models
  • common feature patterns / Common feature patterns
  • good features, creating / Creating good features
• features, dataset
  • URL / More complex pipelines
• feature selection
  • about / Feature selection
  • best individual features, selecting / Selecting the best individual features
• feed-forward neural network
  • about / An introduction to neural networks
• filename, data
  • Blogger ID / Getting the data
  • Gender / Getting the data
  • Age / Getting the data
  • Industry / Getting the data
  • Star Sign / Getting the data
• FP-growth algorithm
  • about / Algorithms for affinity analysis
• frequent itemsets
  • about / Algorithms for affinity analysis
• functions, transformer
  • fit() / The transformer API
  • transform() / The transformer API
• function words
  • about / Function words
  • counting / Counting function words
  • classifying with / Classifying with function words

G

• GPU
  • using, for computation / When to use GPUs for computation
  • benefits / When to use GPUs for computation
  • avenues, defining / When to use GPUs for computation
  • code, running on / Running our code on a GPU
• GPU optimization
  • about / GPU optimization
• graph
  • creating / Creating a graph
• gzip
  • about / Accessing the Enron dataset

H

• Hadoop
  • about / Hadoop MapReduce
  • Distributed File System (HDFS) / Hadoop MapReduce
  • YARN / Hadoop MapReduce
  • Pig / Hadoop MapReduce
  • Hive / Hadoop MapReduce
  • HBase / Hadoop MapReduce
  • courses / Courses on Hadoop
• Hadoop MapReduce
  • about / Hadoop MapReduce
• hash function
  • about / Finding the stories in arbitrary websites
• hidden layer
  • about / An introduction to neural networks
  • creating / An introduction to Lasagne
• hierarchical clustering
  • about / Evidence accumulation

I

• image
  • extracting / Application scenario and goals
• image datasets
  • URL / Mahotas
• input layer
  • about / An introduction to neural networks
• installation instructions, scikit-learn
  • URL / Installing scikit-learn
• instructions, AWS CLI
  • URL / Training on Amazon's EMR infrastructure
• intra-cluster distance
  • about / Optimizing criteria
• Ionosphere
  • about / Loading the dataset
  • URL / Loading the dataset
• Ionosphere Nearest Neighbor
  • about / Loading the dataset
• IPython
  • installing / Installing IPython
  • URL / Installing IPython
• IPython Notebook
  • creating / Downloading data from a social network
  • URL / Extending the IPython Notebook
• IPython notebook
  • using / Using Python and the IPython Notebook
• Iris Setosa / Loading and preparing the dataset
• Iris Versicolour / Loading and preparing the dataset
• Iris Virginica / Loading and preparing the dataset

J

• Jaccard Similarity
  • about / Creating a similarity graph
• JQuery library
  • about / Loading and classifying the dataset
• JSON
  • about / Loading and classifying the dataset
  • and dataset, comparing / Loading and classifying the dataset

K

• k-means algorithm
  • about / The k-means algorithm
  • assignment phase / The k-means algorithm
  • updating phase / The k-means algorithm
• Kaggle
  • URL / More resources
  • about / More resources
• karma
  • about / Reddit as a data source
• Keras
  • URL / Keras and Pylearn2
• kernel
  • about / Loading and classifying the dataset
• kernel parameter
  • about / Kernels
• kernels / Kernels

L

• Lasagne
  • about / An introduction to Lasagne
  • URL / An introduction to Lasagne
• Levenshtein edit distance
  • about / Ranking mechanisms for words
  • computing / Ranking mechanisms for words
• Locality-Sensitive Hashing (LSH)
  • about / Scalability with the nearest neighbor
• local n-grams
  • references / Local n-grams
  • about / Local n-grams
• local optima
  • about / Back propagation
• log probabilities
  • using / Putting it all together

M

• machine-learning workflow
  • training / Testing the algorithm
  • testing / Testing the algorithm
• Mahotas
  • about / Mahotas
  • references / Mahotas
• Manhattan distance
  • about / Distance metrics
• MapReduce
  • about / MapReduce
  • defining / Intuition
  • WordCount example / A word count example
  • Hadoop MapReduce / Hadoop MapReduce
• matplotlib
  • URL / scikit-learn estimators
• MD5 algorithm
  • using / Finding the stories in arbitrary websites
• metadata
  • about / Disambiguation
• MiniBatchKMeans
  • about / Implementation
• Minimum Spanning Tree (MST)
  • about / Evidence accumulation
  • computing / Evidence accumulation
• movie recommendation problem
  • about / The movie recommendation problem
  • dataset, obtaining / Obtaining the dataset
  • loading, with pandas / Loading with pandas
  • sparse data formats / Sparse data formats
• mrjob
  • URL / Training on Amazon's EMR infrastructure
• mrjob package / The mrjob package
• multiple SVMs
  • creating / Classifying with SVMs

N

• n-gram
  • about / Character n-grams
• n-grams
  • about / N-grams
  • disadvantages / N-grams
  • advantages / N-grams
• Naive Bayes
  • about / Naive Bayes
  • Bayes' theorem / Bayes' theorem
  • algorithm / Naive Bayes algorithm
  • working / How it works
• Naive Bayes algorithm
  • mrjob package / The mrjob package
  • blog posts, extracting / Extracting the blog posts
  • Naive Bayes model, training / Training Naive Bayes
  • classifier, running / Putting it all together
  • Amazon's EMR infrastructure, training / Training on Amazon's EMR infrastructure
• Naive Bayes model
  • training / Training Naive Bayes
• NaN (Not a Number)
  • about / Feature creation
• National Basketball Association (NBA)
  • about / Loading the dataset
  • URL / Collecting the data
• Natural Language ToolKit (NLTK)
  • about / Bag-of-words
• nearest neighbor
  • about / scikit-learn estimators
• nearest neighbor algorithm
  • URL / Scalability with the nearest neighbor
• Nearest neighbors
  • about / Nearest neighbors
• network
  • building / Building the network
• networks
  • defining / Deeper networks
• NetworkX
  • URL / Creating a similarity graph, NetworkX
  • defining / NetworkX
• NetworkX package
  • about / Creating a graph
• neural network
  • training / Training and classifying
  • classifying / Training and classifying
  • back propagation (backprop) algorithm / Back propagation
  • words, predicting / Predicting words
• neural network layers, Lasagne
  • network-in-network layers / An introduction to Lasagne
  • dropout layers / An introduction to Lasagne
  • noise layers / An introduction to Lasagne
• Neural networks
  • about / An introduction to neural networks
• neural networks
  • about / scikit-learn estimators, Artificial neural networks, Deep neural networks
  • training / Deep neural networks
  • defining / Intuition
  • implementing / Implementation
  • Theano, defining / An introduction to Theano
  • Lasagne, defining / An introduction to Lasagne
  • implementing, with nolearn / Implementing neural networks with nolearn
  • URL / More resources
• neurons
  • about / Artificial neural networks
• news articles
  • obtaining / Obtaining news articles
  • web API used, for obtaining data / Using a Web API to get data
  • Reddit, as data source / Reddit as a data source
  • data, obtaining / Getting the data
  • clustering / Grouping news articles
  • k-means algorithm / The k-means algorithm
  • results, evaluating / Evaluating the results
  • topic information, extracting from clusters / Extracting topic information from clusters
  • clustering algorithms, using as transformers / Using clustering algorithms as transformers
• NLTK
  • references / Natural language processing and part-of-speech tagging
• NLTK installation instructions
  • URL / Application
• noise
  • adding / Adding noise
• nolearn package
  • neural networks, implementing with / Implementing neural networks with nolearn
• nonprogrammers, for Python language
  • URL / Installing Python
• n_neighbors
  • about / Setting parameters

O

• object classification
  • about / Object classification
• one-versus-all classifier
  • creating / Classifying with SVMs
• OneR
  • about / Implementing the OneR algorithm
• online learning
  • about / Online learning
  • defining / An introduction to online learning
  • implementing / Implementation
• ordinal
  • about / Common feature patterns
• output layer
  • about / An introduction to neural networks
• overfitting
  • about / Testing the algorithm

P

• pagination
  • about / Getting follower information from Twitter
• pandas
  • URL / Collecting the data, More on pandas
  • references / More on pandas
• pandas (Python Data Analysis)
  • about / Collecting the data
• pandas.read_csv function
  • about / Cleaning up the dataset
• pandas documentation
  • URL / Engineering new features
• parameters, ensemble process
  • n_estimators / Parameters in Random forests
  • oob_score / Parameters in Random forests
  • n_jobs / Parameters in Random forests
• petal length / Loading and preparing the dataset
• petal width / Loading and preparing the dataset
• pip
  • about / Installing Python, Creating a graph
• Pipeline
  • creating / Putting it all together
• pipeline
  • creating / Application
  • NLTKBOW transformer / Putting it all together
  • DictVectorizer transformer / Putting it all together
  • BernoulliNB classifier / Putting it all together
• pipelines
  • about / Pipelines
• Pipelines
  • URL / More complex pipelines
• precision
  • about / Evaluation using the F1-score
• preprocessing, using pipelines
  • about / Preprocessing using pipelines
  • features / Preprocessing using pipelines
  • features, of animal / Preprocessing using pipelines
  • example / An example
  • standard preprocessing / Standard preprocessing
  • workflow, creating / Putting it all together
• pricing alerts
  • URL / Training on Amazon's EMR infrastructure
• Principal Component Analysis (PCA)
  • about / Principal Component Analysis
• prior belief
  • about / Bayes' theorem
• probabilistic graphical models
  • URL / More resources
• probabilities
  • computing / Putting it all together
• programmers, for Python language
  • URL / Installing Python
• Project Gutenberg
  • URL / Getting the data
• Pydoop
  • about / Pydoop
  • URL / Pydoop
• Pylearn2
  • about / Keras and Pylearn2
  • URL / Keras and Pylearn2
• Python
  • using / Using Python and the IPython Notebook
  • installing / Installing Python
  • URL / Installing Python
  • defining / Disambiguation
• Python 3.4
  • about / Installing Python

Q

• quotequail package
  • about / Creating a dataset loader

R

• RandomForestClassifier
  • about / Parameters in Random forests
• random forests
  • about / scikit-learn estimators
  • defining / Random forests
  • ensembles, working / How do ensembles work?
  • parameters / Parameters in Random forests
  • applying / Applying Random forests
  • new features, engineering / Engineering new features
• README
  • about / Extracting association rules
• real-time clusterings
  • about / Real-time clusterings
• reasons, feature selection
  • complexity, reducing / Feature selection
  • noise, reducing / Feature selection
  • readable models, creating / Feature selection
• recall
  • about / Evaluation using the F1-score
• recommendation engine
  • building / Recommendation engine
  • URL / Recommendation engine
• reddit
  • about / Obtaining news articles, Using a Web API to get data
  • references / Using a Web API to get data
• Reddit
  • about / Reddit as a data source
  • URL / Reddit as a data source
• regularization
  • URL / Principal Component Analysis
• reinforcement learning
  • URL / Reinforcement learning
• RESTful interface (Representational State Transfer)
  • about / Using a Web API to get data
• rules
  • support / Implementing a simple ranking of rules
  • confidence / Implementing a simple ranking of rules
  • finding / Ranking to find the best rules

S

• sample size
  • increasing / Increasing the sample size
• scikit-learn
  • installing / Installing scikit-learn
  • URL / Installing scikit-learn
• scikit-learn estimators
  • defining / scikit-learn estimators
  • fit() / scikit-learn estimators
  • predict() / scikit-learn estimators
  • Nearest neighbors / Nearest neighbors
  • distance metrics / Distance metrics
  • dataset, loading / Loading the dataset
  • standard workflow, defining / Moving towards a standard workflow
  • fit() function / Moving towards a standard workflow
  • predict() function / Moving towards a standard workflow
  • algorithm, running / Running the algorithm
  • parameters, setting / Setting parameters
• scikit-learn package
  • references / Evaluation
• Scikit-learn tutorials
  • URL / Scikit-learn tutorials
• self-posts
  • about / Reddit as a data source
• sepal length / Loading and preparing the dataset
• sepal width / Loading and preparing the dataset
• shapes adding, CAPTCHAs
  • URL / Better (worse?) CAPTCHAs
• Silhouette Coefficient
  • about / Optimizing criteria
  • computing / Optimizing criteria
  • parameters / Optimizing criteria
• Similarity graph
  • creating / Creating a similarity graph
• SNAP
  • URL / NetworkX
• softmax nonlinearity
  • about / An introduction to Lasagne
• Spam detection
  • references / Spam detection
• spam filter
  • about / Evaluation using the F1-score
• sparse matrix
  • about / Distance metrics
• sparse matrix format
  • about / Sparse data formats
• sports outcome prediction
  • about / Sports outcome prediction
  • features / Sports outcome prediction
• stacking
  • about / Putting it all together
• StackOverflow question
  • URL / More on pandas
• standings
  • loading / Putting it all together
• standings data
  • obtaining / Putting it all together
  • URL / Putting it all together
• Stratified K Fold
  • about / Running the algorithm
• style sheets
  • about / Extracting text from arbitrary websites
• stylometry
  • about / Attributing documents to authors
• subgraphs
  • finding / Finding subgraphs
  • connected components / Connected components
  • criteria, optimizing / Optimizing criteria
• subreddits
  • about / Obtaining news articles, Reddit as a data source
• support / Implementing a simple ranking of rules
• support vector machines (SVM)
  • about / scikit-learn estimators
• SVMs
  • about / Support vector machines
  • URL / Support vector machines
  • classifying with / Classifying with SVMs
  • kernels / Kernels
• system
  • building, for taking image as input / Application scenario and goals

T

• temporal analysis
  • about / Temporal analysis
• text
  • about / Disambiguation
  • extracting, from arbitrary websites / Extracting text from arbitrary websites
• text transformers
  • defining / Text transformers
  • word, counting in dataset / Bag-of-words
  • bag-of-words model / Bag-of-words
  • n-grams / N-grams
  • features / Other features
• tf-idf
  • about / Bag-of-words
• Theano
  • about / An introduction to Theano
  • using / An introduction to Theano
  • URL / Running our code on a GPU
• Torch
  • URL / Keras and Pylearn2
• train_feature_value() function
  • about / Implementing the OneR algorithm
• transformer
  • creating / Creating your own transformer
  • API / The transformer API
  • implementing / Implementation details
  • unit testing / Unit testing
• tutorial, Google
  • URL / Courses on Hadoop
• tutorial, Yahoo
  • URL / Courses on Hadoop
• tweet
  • about / Disambiguation
• tweets
  • loading / Putting it all together
  • F1-score, used for evaluation / Evaluation using the F1-score
  • features, obtaining from models / Getting useful features from models
• Twitter
  • follower information, obtaining from / Getting follower information from Twitter
• Twitter account
  • URL / Downloading data from a social network
• twitter documentation
  • URL / Downloading data from a social network

U

• UCL Machine Learning data repository
  • URL / Loading the dataset
• univariate feature
  • about / Selecting the best individual features
• unstructured format
  • about / Disambiguation
• use cases, computer vision
  • about / Use cases

V

• V's, big data
  • volume / Big data
  • velocity / Big data
  • variety / Big data
  • veracity / Big data
• variance
  • about / How do ensembles work?, Principal Component Analysis
• virtualenv
  • URL / Setting up the environment, Scalability with the nearest neighbor
• vocabulary
  • about / Counting function words
• Vowpal Wabbit
  • about / Vowpal Wabbit
  • URL / Vowpal Wabbit

W

• web-based API, considerations
  • authorization methods / Using a Web API to get data
  • rate limiting / Using a Web API to get data
  • API Endpoints / Using a Web API to get data
• weight
  • about / An introduction to neural networks
• weighted edge
  • about / Creating a similarity graph

Z

• 7-zip
  • URL / Accessing the Enron dataset