Basic modeling course

​

Goals: 

Introductory notions to the use of SPSS Modeler.

​

Techniques presented: 

Introduction to data quality and introduction to data manipulation and data preparation, overview of graphic nodes, advanced data processing topics such as sampling, management of missing data, manipulation of files.

​

Prerequisites: 

Minimum familiarity with the PC, with Windows and with the most common office automation tools is required.

 

Introduction

• Introduction to Modeler

• Methodological outlines and CRISP-DM

• Using the GUI and help

• Modeler Clients and Servers

• Work files, saving and organization in projects

 

Using nodes and analysis

• Reading data files from different sources (db, text, excel, and other proprietary formats)

• Different types of fields, automatic reading and manual setting of the type

• Introduction to data quality:

  • Recognition and treatment of invalids

  • Diagnostic tools to recognize outliers and/or other data anomalies

• Introduction to data manipulation and data preparation

  • Basic operations on records and fields

  • Sorts of fields and records

  • Generation of new fields

  • Use of CLEM language and expression generator

  • Multiple way of creating fields

• Overview of graphical nodes and their use of data analysis/exploration and interactive mode

• Study of univariate distributions and bivariate relationships between data

 

Advanced data processing issues

• Simultaneous use of multiple data sources

  • Aggregations

  • Data source merges

  • Accommodations

• The use of super nodes

• Sampling

  • Basic sampling techniques

  • Sampling for the efficiency of executions

  • Separation of data sources based on samples

  • Data caching

  • Data partitioning

• Management of missing data

  • Setting blank values

  • Using global values for substitutions

  • Automatic check for null and out of range values

  • Advice on managing nulls

• Working with dates

  • General format options and two-digit dates

  • Data readings and transformations

  • Applying formulas to multiple fields

• Working with string data

  • Advanced examples of string manipulation

• Working with sequential data

  • Counts and status

  • Sequential data functions in Modeler

• File manipulation

  • Using aggregation

  • Pure and conditional transpositions of files

  • Creation of flag fields/indicator functions from categorical data

• Notes on the types of joins and SQL optimization

 

Introduction to modelling

• Overview of modeling techniques

  • neural networks

  • Decision trees

  • Statistical prediction (prediction) models

  • Linear regressions and logistic

  • Clustering techniques

  • Dimensionality reduction techniques (principal component analysis)

  • Association rules

• Using a neural network

• Understanding of the logic of neural networks

• Use of decision trees in basic and interactive mode

• Combining and comparing different models

• Example of clustering

• Examples of using association rules