CHAPTER 5.
FORECASTING.
CHAPTER 5.
FORECASTING.
What is this?
Forecasting is the process of making
statements about events whose actual
outcomes have not yet been observed.
CHAPTER 5.
FORECASTING.
What is this? (another definition)
The use of historic data to determine the direction
of future trends. Forecasting is used by companies
to determine how to allocate their budgets for an
upcoming period of time.
The only function of economic forecasting is
to make astrology look respectable.
Ezra Solomon, (1920-2002) Economist
Prediction is difficult, especially if it’s about
the future.
Nil Bohr, (1885-1962) Nobel laureate in
Physics,
"640 kb ought to be enough for anybody." – 1981, Bill
Gates, (1955) (attributed to)
The internet will catastrophically collapse in 1996.
Robert Metcalfe (1946), co-inventor of Ethernet
CHAPTER 5. FORECASTING.
We can be interested in forecasting:
- Lead time.
- Prices.
- Cost.
- Demand.
- …
CHAPTER 5. FORECASTING.
We can be interested in forecasting:
- Lead time.
- Prices.
- Cost.
- Demand.
- …and, why?
CHAPTER 5. FORECASTING.
We can be interested in forecasting:
- Lead time.
- Prices.
- Cost.
- Demand.
- …and, why?
CHAPTER 5. FORECASTING.
Why can we be interested in
forecasting demand? Because
it can be useful for an efficient
utilization of resources.
CHAPTER 5. FORECASTING.
We must admit, however,
that most forecasts will
be erroneus.
CHAPTER 5. FORECASTING.
We must admit, however, that
most forecasts will be
erroneus.(the longer the term,
the worse the forecast).
CHAPTER 5. FORECASTING.
FORECASTING METHODS.
Qualitative Methods:
-Based on judgement, intuition,
experience, surveys.
-Historical data.
CHAPTER 5. FORECASTING.
FORECASTING METHODS.
Qualitative Methods:
Examples: flight tickets demand
gets reduced when there are
public health issues.
CHAPTER 5. FORECASTING.
FORECASTING METHODS.
Quantitative Methods: Time Series Analysis.
- Based on quantitative historical data.
- We can observe some trends, and/or seasonality.
- We expect future behaviour similar to past
behaviour.
CHAPTER 5. FORECASTING.
FORECASTING METHODS.
Quantitative Methods: Time Series
Analysis.
Examples: demand for pianos is
growing at 2,3% annual rate* globally.
([Link]).
CHAPTER 5. FORECASTING.
FORECASTING METHODS.
Quantitative Methods: Causal relationship
forecasting.
- The variable to be estimated as a result of
some specific causes. LINEAR REGRESSION
CHAPTER 5. FORECASTING.
FORECASTING METHODS.
Quantitative Methods: Causal relationship
forecasting.
- The variable to be estimated as a result of some
specific causes. Example: an increase in demand for
brand new cars will result in an increase in demand
for insurance policies and parking lots.
CHAPTER 5. FORECASTING.
FORECASTING METHODS.
Quantitative Methods:
Time Series Analysis and
Causal relationship.
CHAPTER 5. FORECASTING.
FORECASTING
METHODS.
Quantitative Methods:
Time Series Analysis
CHAPTER 5. FORECASTING.
FORECASTING METHODS.
Quantitative Methods: Time
Series Analysis. Some features:
stationary, trend, seasonality,
cycles, and randomness
FORECASTING METHODS.
Quantitative Methods: Time Series Analysis. Some features:
stationary, trend, seasonality, cycles, and randomness. We’ll use
some specific methods:
Stationary time series:
SIMPLE MOVING AVERAGE
(SIMPLE) EXPONENTIAL SMOOTHING
Time series with trend:
DOUBLE EXPONENTIAL SMOOTHING
Time series with seasonality:
Winter Method. (We won’t do it).