Index

A

• A/B testing
  • models, iterating / Iterating on models through A/B testing
  • experimental allocation / Experimental allocation – assigning customers to experiments
  • sample size, deciding / Deciding a sample size
  • multiple hypothesis testing / Multiple hypothesis testing
• adjacency matrix / Where agglomerative clustering fails
• affinity propagation
  • cluster numbers, selecting automatically / Affinity propagation – automatically choosing cluster numbers
• agglomerative clustering
  • about / Agglomerative clustering
  • failures / Where agglomerative clustering fails
• Alternating Least Squares (ALS) / Case Study: Training a Recommender System in PySpark
• Amazon Web Services (AWS) / Working in the cloud
• analytic pipeline
  • data splitting / Modeling layer
  • parameter tuning / Modeling layer
  • model performance / Modeling layer
  • model persistence / Modeling layer
• analytic solution, advanced
  • designing / Designing an advanced analytic solution
  • data layer / Data layer: warehouses, lakes, and streams
  • modeling layer / Modeling layer
  • deployment layer / Deployment layer
  • reporting layer / Reporting layer
• application layer / Deployment layer
• Area Under Curve (AUC) / Evaluating changes in model performance
• area under curve (AUC)
  • about / Evaluating classification models
• auto-regressive moving average (ARMA) / Time series data

B

• back-propagation
  • about / Parameter fitting with back-propagation
• boosting
  • about / Fitting and SVM to the census data, Boosting – combining small models to improve accuracy
• broker / Persisting information with database systems

C

• categorical data
  • similarity metrics / Similarity metrics for categorical data
  • normalizing / Similarity metrics for categorical data
• Celery library
  • URL / The web application
• Classification and Regression Trees (CART) algorithm / Decision trees
• classification models
  • evaluating / Evaluating classification models
  • improving / Strategies for improving classification models
• client layer / Deployment layer
• client requests
  • handling / Clients and making requests
  • GET requests, implementing / The GET requests
  • POST request, implementing / The POST request
  • HEAD request, implementing / The HEAD request
  • PUT request, implementing / The PUT request
  • DELETE request, implementing / The DELETE request
• communication
  • guidelines / Guidelines for communication
  • terms, translating to business values / Translate terms to business values
  • results, visualizing / Visualizing results
• convexity
  • about / Jointly optimizing all parameters with second-order methods
• convolutional network
  • about / Convolutional networks and rectified units
  • input layer / Convolutional networks and rectified units
  • convolutional layer / Convolutional networks and rectified units
  • rectifying layer / Convolutional networks and rectified units
  • downsampling layer / Convolutional networks and rectified units
  • fully connected layer / Convolutional networks and rectified units
• correlation similarity metrics
  • about / Correlation similarity metrics and time series
• covariance / Correlation similarity metrics and time series
• curl command
  • about / The architecture of a prediction service
  • URL / The architecture of a prediction service

D

• database systems
  • using / Persisting information with database systems
• data layer / Designing an advanced analytic solution
• decision trees
  • about / Decision trees
• dendrograms / Agglomerative clustering
• deployment layer / Deployment layer
• digit recognition / The TensorFlow library and digit recognition
• distance metrics
  • about / Similarity and distance metrics
  • numerical distance metrics / Numerical distance metrics
  • time series / Correlation similarity metrics and time series
  • blending / Similarity metrics for categorical data
• Dow Jones Industrial Average (DJIA) / Correlation similarity metrics and time series
• Driver / Creating the SparkContext
• Dynamic Time Warping (DTW) / Correlation similarity metrics and time series

E

• e-mail campaigns, case study
  • about / Case study: targeted e-mail campaigns
  • data input and transformation / Data input and transformation
  • sanity checking / Sanity checking
  • model development / Model development
  • scoring / Scoring
  • visualization and reporting / Visualization and reporting
• Executors / Creating the SparkContext

F

• false positive rate (FPR)
  • about / Evaluating classification models
• familywise error rate (FWER) / Multiple hypothesis testing
• Flask
  • URL / Application – the engine of the predictive services

G

• Gaussian kernel
  • about / Fitting and SVM to the census data
• Gauss Markov Theorem / Linear regression
• generalized linear models
  • about / Generalized linear models
• Generalized Linear Models (GLMs) / Logistic regression
• Generalize Estimating Equations (GEE)
  • about / Generalize estimating equations
• geospatial data
  • about / Working with geospatial data
  • loading / Loading geospatial data
  • cloud, working in / Working in the cloud
• gradient boosted decision trees
  • about / Gradient boosted decision trees
  • versus, support vector machines and logistic regression / Comparing classification methods
• gradient boosted machine (GBM) / Evaluating changes in model performance
• graphical user interface (GUI) / Cleaning textual data
• graphics processing unit (GPU) / The TensorFlow library and digit recognition

H

• H20
  • URL / Joining signals and correlation
• Hadoop distributed file system (HDFS) / Creating an RDD
• hierarchical clustering / Agglomerative clustering
• hinge loss
  • about / Separating Nonlinear boundaries with Support vector machines
• horizontal scaling / Server – the web traffic controller
• HTTP Status Codes / The GET requests
• hypertext transfer protocol (HTTP)
  • about / The architecture of a prediction service

I

• images
  • about / Images
  • image data, cleaning / Cleaning image data
  • thresholding, for highlighting objects / Thresholding images to highlight objects
  • dimensionality reduction, for image analysis / Dimensionality reduction for image analysis
• Indicator Function / Extracting features from textual data
• Internet Movie Database
  • URL / Exploring categorical and numerical data in IPython
• IPython notebook
  • about / Exploring categorical and numerical data in IPython
  • installing / Installing IPython notebook
  • interface / The notebook interface
  • data, loading / Loading and inspecting data
  • data, inspecting / Loading and inspecting data
  • basic manipulations / Basic manipulations – grouping, filtering, mapping, and pivoting
  • Matplotlib, charting with / Charting with Matplotlib
• iteratively reweighted least squares (IRLS)
  • about / Jointly optimizing all parameters with second-order methods

K

• K-means ++ / K-means clustering
• K-means clustering
  • about / K-means clustering
• k-medoids
  • about / k-medoids
• kernel function
  • about / Separating Nonlinear boundaries with Support vector machines

L

• Labeled RDD / Streaming clustering in Spark
• Latent Dirichlet Allocation (LDA)
  • about / Latent Dirichlet Allocation
• Latent Semantic Indexing (LSI) / Principal component analysis
• linear regression
  • about / Linear regression
  • data, preparing / Data preparation
  • evaluation / Model fitting and evaluation
  • model, fitting / Model fitting and evaluation
  • statistical significance / Statistical significance of regression outputs
  • Generalize Estimating Equations (GEE) / Generalize estimating equations
  • mixed effects models / Mixed effects models
  • time series data / Time series data
  • generalized linear models / Generalized linear models
  • regularization, applying to linear models / Applying regularization to linear models
• linkage metric / Where agglomerative clustering fails
• link functions
  • Logit / Generalized linear models
  • Poisson / Generalized linear models
  • Exponential / Generalized linear models
• logistic regression
  • about / Logistic regression
  • multiclass logistic classifiers / Multiclass logistic classifiers: multinomial regression
  • dataset, formatting for classification problems / Formatting a dataset for classification problems
  • stochastic gradient descent (SGD) / Learning pointwise updates with stochastic gradient descent
  • parameters, optimizing with second-order methods / Jointly optimizing all parameters with second-order methods
  • model, fitting / Fitting the model
  • versus, support vector machines and gradient boosted decision trees / Comparing classification methods
• logistic regression service
  • as case study / Case study – logistic regression service
  • database, setting up / Setting up the database
  • web server, setting up / The web server
  • web application, setting up / The web application
  • model, training / The flow of a prediction service – training a model
  • on-demand and bulk prediction, obtaining / On-demand and bulk prediction
• Long Short Term Memory Networks (LSTM) / Optimizing the learning rate

M

• Matplotlib
  • charting with / Charting with Matplotlib
• message passing / Affinity propagation – automatically choosing cluster numbers
• Mixed National Institute of Standards and Technology (MNIST) database / The MNIST data
• modeling layer / Modeling layer
• model performance
  • checking, with diagnostic / Checking the health of models with diagnostics
  • changes, evaluating / Evaluating changes in model performance
  • changes in feature importance, evaluating / Changes in feature importance
  • unsupervised model performance, changes / Changes in unsupervised model performance
• models
  • iterating, through A/B testing / Iterating on models through A/B testing
• multiclass logistic classifiers
  • about / Multiclass logistic classifiers: multinomial regression
• multidimensional scaling (MDS) / Numerical distance metrics
• multinomial regression / Multiclass logistic classifiers: multinomial regression

N

• natural language toolkit (NLTK) library / Cleaning textual data
• neural networks
  • patterns, learning with / Learning patterns with neural networks
  • perceptron / A network of one – the perceptron
  • perceptrons, combining / Combining perceptrons – a single-layer neural network
  • single-layer neural network / Combining perceptrons – a single-layer neural network
  • parameter fitting, with back-propagation / Parameter fitting with back-propagation
  • discriminative, versus generative models / Discriminative versus generative models
  • gradients, vanishing / Vanishing gradients and explaining away
  • belief networks, pretraining / Pretraining belief networks
  • regularizing, dropout used / Using dropout to regularize networks
  • convolutional networks / Convolutional networks and rectified units
  • rectified units / Convolutional networks and rectified units
  • data compressing, with autoencoder networks / Compressing Data with autoencoder networks
  • learning rate, optimizing / Optimizing the learning rate
• neurons / Combining perceptrons – a single-layer neural network
• Newton methods
  • about / Jointly optimizing all parameters with second-order methods
• non-relational database / Persisting information with database systems
• numerical distance metrics
  • about / Numerical distance metrics

O

• Ordinary Least Squares (OLS) / Linear regression

P

• prediction service
  • architecture / The architecture of a prediction service
  • sever, using / Server – the web traffic controller
  • application, setting up / Application – the engine of the predictive services
  • information, persisting with database systems / Persisting information with database systems
• Principal Component Analysis (PCA)
  • about / Principal component analysis
  • Latent Dirichlet Allocation (LDA) / Latent Dirichlet Allocation
  • dimensionality reduction, using in predective modeling / Using dimensionality reduction in predictive modeling
• pseudo-residuals / Gradient boosted decision trees
• pyspark
  • classifier models, implementing / Case study: fitting classifier models in pyspark
• PySpark
  • URL / Joining signals and correlation, Introduction to PySpark
  • about / Introduction to PySpark, Scaling out with PySpark – predicting year of song release
  • SparkContext, creating / Creating the SparkContext
  • RDD, creating / Creating an RDD
  • Spark DataFrame, creating / Creating a Spark DataFrame
  • example / Scaling out with PySpark – predicting year of song release
• Python requests library
  • URL / The GET requests

R

• RabbitMQ
  • URL / The web application
• random forest
  • about / Random forest
• RDD
  • creating / Creating an RDD
• Receiver-Operator-Characteristic (ROC) / Evaluating changes in model performance
• receiver operator characteristic (ROC) / Logistic regression
• Receiver Operator Characteristic (ROC) curve
  • about / Evaluating classification models
• recommender system training, in PySpark
  • case study / Case Study: Training a Recommender System in PySpark
• Rectified Linear Unit (ReLU) / Convolutional networks and rectified units
• Recurrent Neural Networks (RNNs) / Optimizing the learning rate
• Redis
  • URL / Setting up the database
• relational database / Persisting information with database systems
• reporting layer / Reporting layer
• reporting service
  • about / Case Study: building a reporting service
  • report server, setting up / The report server
  • report application, setting up / The report application
  • visualization layer, using / The visualization layer
• Resilient Distributed Dataset (RDD) / Streaming clustering in Spark
• Resilient Distributed Datasets (RDDs) / Introduction to PySpark

S

• second-order methods
  • about / Formatting a dataset for classification problems
  • parameters, optimizing / Jointly optimizing all parameters with second-order methods
• server
  • used, for communicating with external systems / Server – the web traffic controller
• similarity metrics
  • about / Similarity and distance metrics
  • correlation similarity metrics / Correlation similarity metrics and time series
  • for categorical data / Similarity metrics for categorical data
• Singular Value Decomposition (SVD) / Numerical distance metrics, Principal component analysis
• social media feeds, case study
  • about / Case study: sentiment analysis of social media feeds
  • data input and transformation / Data input and transformation
  • sanity checking / Sanity checking
  • model development / Model development
  • scoring / Scoring
  • visualization and reporting / Visualization and reporting
• soft-margin formulation / Separating Nonlinear boundaries with Support vector machines
• Spark
  • streaming clustering / Streaming clustering in Spark
• SparkContext
  • creating / Creating the SparkContext
• Spark DataFrame
  • creating / Creating a Spark DataFrame
• spectral clustering / Where agglomerative clustering fails
• statsmodels
  • URL / Model fitting and evaluation
• stochastic gradient descent
  • about / Learning pointwise updates with stochastic gradient descent
• stochastic gradient descent (SGD)
  • about / Formatting a dataset for classification problems
• streaming clustering
  • about / Streaming clustering in Spark
• support-vector networks
  • about / Separating Nonlinear boundaries with Support vector machines
• support vector machine (SVM)
  • nonlinear boundaries, separating / Separating Nonlinear boundaries with Support vector machines
  • implementing, to census data / Fitting and SVM to the census data
  • boosting / Boosting – combining small models to improve accuracy
  • versus, logistic regression and gradient boosted decision trees / Comparing classification methods

T

• TensorFlow library
  • about / The TensorFlow library and digit recognition
  • MNIST data / The MNIST data
  • network, constructing / Constructing the network
• term-frequency-inverse document frequency (tf-idf) / Extracting features from textual data
• textual data
  • working with / Working with textual data
  • cleaning / Cleaning textual data
  • features, extracting from / Extracting features from textual data
  • dimensionality reduction, used for simplyfying datasets / Using dimensionality reduction to simplify datasets
• time series
  • about / Correlation similarity metrics and time series
• time series analysis
  • about / Time series analysis
  • cleaning and converting / Cleaning and converting
  • time series diagnostics / Time series diagnostics
  • signals and correlation, joining / Joining signals and correlation
• transformations and operations
  • URL / Creating an RDD
• tree methods
  • about / Tree methods
  • decision trees / Decision trees
  • random forest / Random forest
• true positive rate (TPR)
  • about / Evaluating classification models

U

• units / Combining perceptrons – a single-layer neural network
• Unweighted Pair Group Method with Arithmetic Mean (UPGMA) / Agglomerative clustering

V

• vertical scaling / Server – the web traffic controller

W

• Web Server Gateway Interface (WSGI)
  • about / The architecture of a prediction service

X

• XGBoost
  • URL / Joining signals and correlation