• This field is for validation purposes and should be left unchanged.
analytics program enablement for growth-mindset organizations

PredictiveInsight™ Training Outline

(Affinium Model 7.3 and 5.5 also supported)

DAY 1 MORNING

1) Basics of Predictive Modeling

a) What is Modeling?
b) Stages in the modeling process (CRISP-DM)
i) Business Understanding
ii) Data Understanding
(1) Collect data
(2) Describe and explore data
(3) Assess data quality
iii) Data Preparation
(1) Clean data
(a) Missing values
(b) Miscoded data
(2) Construct data
(3) Sample data
iv) Modeling
(1) Select modeling techniques
(2) Build model
(3) Assess model
v) Evaluation
(1) Evaluate Results
(a) Model accuracy
(b) Model interpretation
(2) Review Modeling Process
(3) Accept or Reject Model
vi) Deployment
(1) Determine deployment method
(2) Devise model maintenance plan

2) A simple modeling example
a) Layout of PredictiveInsightTM
b) The Data Import Wizard/ Importing the vetresp.dat file
c) The Modeling Wizard/ building the Quick Model
i) Specifying the response variable
ii) Selecting input variables
iii) Importing the data dictionary
iv) Selecting the modeling level

DAY 1- AFTERNOON

d) Viewing sample reports
i) The Data Dictionary report
ii) The Variable Summary report
iii) The Variable Numeric report
iv) The Variable Profile report
v) The Modeling Summary report
vi) The Model Sensitivity Summary report
vii) The Model Variable Sensitivity report
viii) The Model Performance report
ix) The Campaign report
x) The Model Details report
xi) The Log report

3) A more detailed view
a) Data types
i) Money
ii) Date
iii) Time
iv) Telephone/Access #
v) Flag
vi) Categorical
vii) Quantity
viii) Descriptive/Names
ix) City
x) Zip Code
xi) Country
xii) Continent
xiii) Time Zone
xiv) As Is
b) Data preparation- data cleanup function
c) PredictiveInsight and data preprocessing
i) Money- log ratio preprocessing
ii) Ordered numeric variables
(1) Polynomial
(2) Unsorted chi-squared binning
(3) Z-score normalization
iii) Categorical variables- chi-squared binning

DAY 2- MORNING

d) Algorithm overview
i) RFM
ii) Bayes
iii) Linear regression
iv) Logistic regression
v) Backpropogation neural network
vi) ChAID
vii) CART
viii) Manual

4) Scoring Models
a) Scoring options
b) Deployment options
c) Understanding reports and customizing

5) Other modules
a) Customer Valuator
b) Cross Seller
c) Customer Segmenter

DAY 2- AFTERNOON

6) Client data and projects, questions and answers

 

COURSE MATERIAL

A comprehensive binder of course notes is included for up to 8 participants.  You may purchase additional or replacement sets of the course book here.

 

<< Return to Main Unica Services Page >>