Guolin Lai

DSC8240 Course Web

Objective Hierarchies
Variables and Attributes
Influence Diagram
Mathematical Representation
Testing and Validation
Implementation and Use

Objective Hierarchies

Demand forecasting plays a crucial role in making the decision of a firm's manufacturing or retail store's order.

Such techniques as regression analysis, and time series forecasting (StatsPro) were employed.

Variables and Attributes

AFD include:

• Exogenous Demand: Macro-Economic Influences
• Endogenous Demand: Industry Behavior

• Actual average industry demand
• Average price
• Average advertising
• Average advertising one quarter ago
• Average advertising two quarters ago
• Average R&D expenditures one quarter ago
• Average R&D expenditures two quarters ago

NSOM include:

• Pricing
• Promotion
• Quality
• Loyalty

Inputs for NSOM:

• Actual normalized share of market
• Relative price
• Relative advertising expenditure
• Relative advertising expenditure one quarter ago
• Relative advertising expenditure two quarters ago
• Relative R&D expenditure one quarter ago
• Relative R&D expenditure two quarters ago
• Normalized share of market

Influence Diagram

Mathematical Representation

• stimate Exogenous Demand using time series analysis
• Estimate Endogenous Demand using regression analysis
• Estimate NSOM using regression analysis = Firm Demand / Avg Firm Demand
• Combine above two to get AFD (AFD = Exogenous Demand + Endogenous Demand)
• Combine AFD and NSOM ( Firm Demand = AFD * NSOM = {(T*S) + (B0 + B1*Avg Pricing + B2*Avg Advertising + …)
• Exogenous demand = "Base demand" * Seasonal effects where estimation of base demand and seasonal demand is done using Time Series Decomposition
• Endogenous demand = Influence of aggregate industry behavior

Testing and Validation

Firm Demand Excel Files

Implementation and Use

The model can be manipulated in Microsoft Excel.