Page 68 - CITP Review

Types

Predictive models
Predictive models are used to analyze the relationship between the specific performance of a unit in a
sample and one or more known attributes or features of the unit. The goal of such a model is to
14
determine whether a similar unit in another sample would be likely to exhibit the same attribute. This
category encompasses, for example, fraud detection models. Predictive models may be used in real time
to help with decision making, such as in risk evaluation of a particular course of action.

Descriptive models

Descriptive models are often used to sort customers or users into groups by quantifying relationships in
15
data. Whereas predictive models focus on single behaviors or attributes, descriptive models can track
multiple attributes and relationships. Descriptive models are generally used for categorizing customers
by certain attributes rather than predicting the likelihood of their actions, as predictive models do.
However, descriptive modeling tools can be used to create additional models that in turn can be used to
predict behavior.

Decision models
Decision models can help predict the results of decisions that involve numerous variables by mapping
the relationships of all decision elements. Organizations can use these models maximize or minimize
outcomes, thereby optimizing the decision process. Decision models are most often used to form
decision logic or rules to produce a desired decision outcome for every customer or user.

Techniques
The following sections describe various techniques available for predictive analysis.

Regression
Regression is a statistical tool for measuring the relation between variables. In business, regression is
generally used either for forecasting or optimization. Regressions range from simple models to highly
complex equations. Regressions can range anywhere from simple models to extremely complex models
16
and equations. Simple regression can be used to better understand the relationship between a single
dependent and a single independent variable; the regression results can be used to predict the dependent
17
variable when the independent variable is known.

14
See https://www.slideshare.net/JeffreyStricklandPhD/predictive-modeling-and-analytics-selectchapters-
41304405. Last accessed September 10, 2019.
15
See https://www.slideshare.net/JeffreyStricklandPhD/predictive-modeling-and-analytics-selectchapters-
41304405. Last accessed September 10, 2019.
16
See https://smallbusiness.chron.com/application-regression-analysis-business-77200.html. Last accessed
September 10, 2019.
17
See https://www.statpac.com/statistics-calculator/correlation-regression.htm. Last accessed September 10,
2019.

© 2019 Association of International Certified Professional Accountants. All rights reserved. 2-22