ModelDriven Business Intelligence Systems: Part II

1
Model-Driven Business Intelligence Systems Part
II

• Week 9
• Dr. Jocelyn San PedroSchool of Information
  Management Systems
• Monash University

2
Lecture Outline

• Trend Analysis
• Seasonality Analysis
• Multiplicative Decomposition of a Time Series
• Causal Forecasting Models
• Decision Trees
• Influence Diagrams

3
Learning Objectives

• At the end of this lecture, the students will
• Have understanding of some models used in
  model-driven business intelligence systems
• Specifically, have understanding of trend
  analysis, and seasonality analysis decision
  trees and influence diagrams for decision
  modelling

4
Trend Analysis

• Fits a trend equation (or curve) to a series of
  historical data points
• Projects this curve into the future for medium-
  and long-term forecasts
• Trend equations linear, quadratic, exponential,
  

5
Linear Regression

• Least Squares Procedure
• Fits a line that minimises the sum of the squares
  of vertical differences from the line to each of
  the actual observations i.e. minimises the sum
  of squared errors
• Least squares line Y a bX
• a is the y-axis intercept
• b is the slope of the regression line

6
(No Transcript)
7
Linear Trend Analysis- ExcelModules
8
Seasonality Analysis

• Recurring variations at certain periods (i.e.,
  months) of the year make a seasonal adjustment in
  the time series necessary
• E.g., demand for coal and oil fuel usually peaks
  in cold winter months demand for sunscreen may
  be highest in summer
• Seasonal Index ratio of the average value of
  the item in season to the overall annual average
  value

9
Example - ExcelModules
10
Seasonality Analysis

• Seasonal Index lt1 indicates demand is below
  average that month
• Seasonal index gt1 indicated demand is above
  average that month
• Use the seasonal indices to adjust the monthly
  demand for any future month
• Example If 3rd years average demand is 100
  units,
• forecast for Januarys monthly demand is 100 x
  0.957 96 units, (which is below average)
• Forecast for Mays monthly demand is 100 x 1.309
  131 units, (which is above average)

11
Multiplicative Decomposition of a Time Series

• Breaks down a time series into two components
• Seasonal component
• A combination of the trend and cycle component
  (simply called trend)
• Forecast is calculated a product of composite
  trend and seasonality components

12
Multiplicative Decomposition in ExcelModules
13
Causal Forecasting Models

• Purpose is to develop a mathematical relationship
  between one or more factors affecting a variable
• Example sales of swimwear are likely to depend
  on average daily temperature, price, advertising
  budget
• Sales dependent variable
• average daily temperature, price, advertising
  budget independent variables
• Most common methods
• Linear regression Y a bX
• Multiple regression Y ab1X1b2X2 bpXp

14
Influence diagrams

• An influence diagram is a simple visual
  representation of a decision problem
• Influence diagrams offer an intuitive way to
  identify and display the essential elements,
  including decisions, uncertainties, and
  objectives, and how they influence each other.
• http//www.lumina.com/software/influencediagrams.h
  tml

15
Influence Diagrams
http//www.lumina.com/software/influencediagrams.h
tml
16
http//www.lumina.com/software/influencediagrams.h
tml
17
Example
Influence diagram for RD and commercialization
of a new product
http//www.lumina.com/software/influencediagrams.h
tml
18
Example - Genie
http//www2.sis.pitt.edu/genie/
19
Example - Genie
http//www2.sis.pitt.edu/genie/
20
Decision Trees
http//www.lumina.com/software/influencediagrams.h
tml
21
Example TreePlan
Render, B., Stair, R. and Balakrishnan, N.
(2003) Managerial Decision Modeling, Prentice
Hall.
22
References

• Langley, R. (1970) Practical Statistics Simple
  Explained, Dover Publications, NY.
• Render, B., Stair, R. and Balakrishnan, N. (2003)
  Managerial Decision Modeling, Prentice Hall.
• Render, B., and Stair, R. (1999) Quantitative
  Analysis for Management (or any edition)
• Rowntree, D. (1981) Statistics Without Tears A
  Primer for Non-mathematicians, Penguin Books.
• Useful online resources
• Analytica www.lumina.com/software/influencediagram
  s.html
• Genie - www2.sis.pitt.edu/genie/

23

• Questions?
• Jocelyn.sanpedro_at_sims.monash.edu.au
• School of Information Management and Systems,
  Monash University
• T1.28, T Block, Caulfield Campus
• 9903 2735