IBM SPSS Modeler Cookbook电子书

IBM SPSS Modeler Cookbook

Table of Contents

IBM SPSS Modeler Cookbook

Credits

Foreword

About the Authors

About the Reviewers

www.PacktPub.com

Support files, eBooks, discount offers, and more

Why Subscribe?

Free Access for Packt account holders

Instant Updates on New Packt Books

Preface

What is CRISP-DM?

Data mining is a business process

The IBM SPSS Modeler workbench

A brief history of the Clementine workbench

Historical introduction to scripting

What this book covers

Who this book is for

Conventions

Reader feedback

Customer support

Downloading the example code

Errata

Piracy

Questions

1. Data Understanding

Introduction

Using an empty aggregate to evaluate sample size

Getting ready

How to do it...

How it works...

There's more...

A modified version

See also

Evaluating the need to sample from the initial data

Getting ready

How to do it...

How it works...

There's more...

See also

Using CHAID stumps when interviewing an SME

Getting ready

How to do it...

How it works...

See also

Using a single cluster K-means as an alternative to anomaly detection

Getting ready

How to do it...

How it works...

There's more...

Using an @NULL multiple Derive to explore missing data

Getting ready

How to do it...

How it works...

See also

Creating an Outlier report to give to SMEs

Getting ready

How to do it...

How it works...

See also

Detecting potential model instability early using the Partition node and Feature Selection node

Getting ready

How to do it...

How it works...

See also

2. Data Preparation – Select

Introduction

Using the Feature Selection node creatively to remove or decapitate perfect predictors

Getting ready

How to do it...

How it works...

There's more...

See also

Running a Statistics node on anti-join to evaluate the potential missing data

Getting ready

How to do it...

How it works...

See also

Evaluating the use of sampling for speed

Getting ready

How to do it...

How it works...

There's more...

See also

Removing redundant variables using correlation matrices

Getting ready

How to do it...

How it works...

There's more...

See also

Selecting variables using the CHAID Modeling node

Getting ready

How to do it...

How it works...

There's more...

See also

Selecting variables using the Means node

Getting ready

How to do it...

How it works...

There's more...

See also

Selecting variables using single-antecedent Association Rules

Getting ready

How to do it...

How it works...

There's more...

See also

3. Data Preparation – Clean

Introduction

Binning scale variables to address missing data

Getting ready

How to do it...

How it works...

See also

Using a full data model/partial data model approach to address missing data

Getting ready

How to do it...

How it works...

There's more...

See also

Imputing in-stream mean or median

Getting ready

How to do it...

How it works...

There's more...

See also

Imputing missing values randomly from uniform or normal distributions

Getting ready

How to do it...

How it works...

There's more...

See also

Using random imputation to match a variable's distribution

Getting ready

How to do it...

How it works...

There's more...

See also

Searching for similar records using a Neural Network for inexact matching

Getting ready

How to do it...

How it works...

There's more...

See also

Using neuro-fuzzy searching to find similar names

Getting ready

How to do it...

How it works...

There's more...

See also

Producing longer Soundex codes

Getting ready

How to do it...

How it works...

There's more...

See also

4. Data Preparation – Construct

Introduction

Building transformations with multiple Derive nodes

Getting ready

How to do it...

How it works...

There's more...

Calculating and comparing conversion rates

Getting ready

How to do it...

How it works...

There's more...

See also

Grouping categorical values

Getting ready

How to do it...

How it works...

There's more...

Transforming high skew and kurtosis variables with a multiple Derive node

Getting ready

How to do it...

How it works...

There's more...

Creating flag variables for aggregation

Getting ready

How to do it...

How it works...

There's more...

Using Association Rules for interaction detection/feature creation

Getting ready

How to do it...

How it works...

There's more...

Creating time-aligned cohorts

Getting ready

How to do it...

How it works...

There's more...

5. Data Preparation – Integrate and Format

Introduction

Speeding up merge with caching and optimization settings

Getting ready

How to do it...

How it works...

See also

Merging a lookup table

Getting ready

How to do it...

How it works...

See also

Shuffle-down (nonstandard aggregation)

Getting ready

How to do it...

How it works...

There's more...

See also

Cartesian product merge using key-less merge by key

Getting ready

How to do it...

How it works...

There's more...

See also

Multiplying out using Cartesian product merge, user source, and derive dummy

Getting ready

How to do it...

How it works...

There's more...

See also

Changing large numbers of variable names without scripting

Getting ready

How to do it...

How it works...

There's more...

See also

Parsing nonstandard dates

Getting ready

How to do it...

How it works...

There's more...

Nesting functions into one Derive node

Performing clean downstream of a calculation using a Filter node

Using parameters instead of constants in calculations

See also

Parsing and performing a conversion on a complex stream

Getting ready

How to do it...

How it works...

See also

Sequence processing

Getting ready

How to do it...

How it works...

There's more...

See also

6. Selecting and Building a Model

Introduction

Evaluating balancing with Auto Classifier

Getting ready

How to do it...

How it works...

See also

Building models with and without outliers

Getting ready

How to do it...

How it works...

See also

Using Neural Network for Feature Selection

Getting ready

How to do it...

How it works...

There's more...

See also

Creating a bootstrap sample

Getting ready

How to do it...

How it works...

There's more...

See also

Creating bagged logistic regression models

Getting ready

How to do it...

How it works...

There's more...

See also

Using KNN to match similar cases

Getting ready

How to do it...

How it works...

See also

Using Auto Classifier to tune models

Getting ready

How to do it...

How it works...

See also

Next-Best-Offer for large datasets

Getting ready

How to do it...

How it works...

There's more...

See also

7. Modeling – Assessment, Evaluation, Deployment, and Monitoring

Introduction

How (and why) to validate as well as test

Getting ready

How to do it...

How it works...

See also

Using classification trees to explore the predictions of a Neural Network

Getting ready

How to do it...

How it works...

See also

Correcting a confusion matrix for an imbalanced target variable by incorporating priors

Getting ready

How to do it...

How it works...

There's more...

See also

Using aggregate to write cluster centers to Excel for conditional formatting

Getting ready

How to do it...

How it works...

See also

Creating a classification tree financial summary using aggregate and an Excel Export node

Getting ready

How to do it...

How it works...

See also

Reformatting data for reporting with a Transpose node

Getting ready

How to do it...

How it works...

There's more...

See also

Changing formatting of fields in a Table node

Getting ready

How to do it...

How it works...

There's more...

See also

Combining generated filters

Getting ready

How to do it...

How it works...

There's more...

See also

8. CLEM Scripting

Introduction

CLEM scripting best practices

CLEM scripting shortcomings

Building iterative Neural Network forecasts

Getting ready

How to do it...

How it works...

Script section 1

There's more...

Quantifying variable importance with Monte Carlo simulation

Getting ready

How to do it...

How it works...

Script section 1

Script section 2

There's more...

Implementing champion/challenger model management

Getting ready

How to do it...

How it works...

Script section 1

Script section 2

There's more...

Detecting outliers with the jackknife method

Getting ready

How to do it...

How it works...

Script section 1

Script section 2

Script section 3

There's more...

Optimizing K-means cluster solutions

Getting ready

How to do it...

How it works...

Script section 1

Script section 2

Script section 3

Script section 4

There's more...

Automating time series forecasts

Getting ready

How to do it...

How it works...

Script section 1

Script section 2

There's more...

Automating HTML reports and graphs

Getting ready

How to do it...

How it works...

Script section 1

Script section 2

Script section 3

There's more...

Rolling your own modeling algorithm – Weibull analysis

Getting ready

How to do it...

How it works...

Script section 1

There's more...

A. Business Understanding

Introduction

What decisions are you trying to make using data?

Define business objectives by Tom Khabaza

The importance of business objectives in data mining

Defining the business objectives of a data mining project

Understanding the goals of the business

Understanding the objectives of your client

Connecting specific objectives to analytical results

Specifying your data mining goals

Assessing the situation by Meta Brown

Taking inventory of resources

Reviewing requirements, assumptions, and constraints

Identifying risks and defining contingencies

Defining terminology

Evaluating costs and benefits

Translating your business objective into a data mining objective by Dean Abbott

The key to the translation – specifying target variables

Data mining success criteria – measuring how good the models actually are

Success criteria for classification

Success criteria for estimation

Other customized success criteria

Produce a project plan – ensuring a realistic timeline by Keith McCormick

Business understanding

Data understanding

Data preparation

Modeling

Evaluation

Deployment

Index