week three Flashcards
forecasting
A forecast is:
1. Only as good as the information included in the forecast.
2. Never perfect (due to uncertainty)
unsertency
Uncertainty is “the difference between the amount of information
required and the amount already possessed by the organization”
ou should always ask yourself the following question:
To what extent is the past related to what you expect to see in the future
types of forcasting methodes
Quantitative methods: rely on quantitative data and analytical
techniques.
Qualitative methods: based on subjective opinions from one or
more experts
time series + examples
A variable that is measured over time in sequential order is called a
time series.
We analyze time series to detect patterns in order to forecast the
future values of the time series.
Type of patterns:
1. Trend
2. Seasonal variation
3. Cyclical variation
4. Random variation
- trent
A trend is a long-term relatively smooth pattern or direction, that
persists usually for more than one year
–> What products can you think of that behave this way?
Electric Vehicles (EVs): Demand for electric vehicles is steadily rising due to environmental concerns and advancements in battery technology.
Organic Food Products: Sales of organic food products have shown a steady upward trend as consumers become more health-conscious.
Streaming Services: Subscriptions to streaming services like Netflix and Spotify have grown consistently over recent years as people shift away from cable TV and physical media
- seaonal variation
The seasonal component of the time series exhibits a short-term
(less than one year) calendar repetitive behavior.
–> What products can you think of that behave this way?
————–
Sunscreen: Demand for sunscreen spikes in summer due to increased outdoor activities and sunny weather, then drops during colder months.
Back-to-School Supplies: Items like notebooks, pens, and backpacks see a surge in demand in late summer before the new school year begins.
Holiday Decorations: Sales of Christmas trees, ornaments, and Halloween costumes are highly seasonal, with demand peaking just before each holiday and dropping afterward.
- cyclical variation
A cycle is a wavelike pattern describing a long-term behavior
(for more than one year).
–> What products can you think of that behave this way?
————————–
Luxury Cars: Demand for luxury cars often rises during economic booms when people feel more financially secure and falls during recessions.
Real Estate: Sales of home improvement products (e.g., paint, flooring) tend to increase during economic expansions when more people buy homes and renovate, and decrease during downturns.
High-End Electronics: Expensive electronics, like high-end TVs or gaming consoles, often follow economic cycles, with higher demand during prosperous times and lower demand during recessions.
- random variation
Random variation comprises the irregular unpredictable changes in
the time series. It tends to hide the other (more predictable)
components.
–> What products can you think of that behave this way?
————————
Face Masks and Hand Sanitizers: Demand for these products surged unexpectedly in early 2020 due to the COVID-19 pandemic, a random event that couldn’t have been forecasted in advance.
Bottled Water: Demand for bottled water can spike unexpectedly in areas affected by natural disasters, such as hurricanes or floods, when access to clean water becomes limited.
Generators: Sales of generators can surge after unexpected power outages or extreme weather events, like snowstorms, that disrupt electricity supplies.
forecasting methodes1. moving average
The moving average model uses the last t periods to
predict demand in period t+1.
There can be two types of moving average models:
a) Simple Moving Average (SMA)
b) Weighted Moving Average (WMA)
* Assumption: the most accurate prediction of future
demand is a simple (linear) combination of past demand.
simple moving average (SMA)
see papier
but the wrong in this is–Because we used equal weights, the downward trend
that actually exists is not observed…
forecast methode 2 weighted moving average (WMA)
sea paper
What if we use a 6-month weighted
moving average?
Weight for last three months more
than the first three months…
Use the following three set of weights:
1. 20/80%
2. 30/70%
3. 40/60%
The more weight is given to recent data,
the more we pick up the declining trend in our forecast.
NOTE: The weight is chosen based on the importance we feel the past data has
forcasting methode Exponential smoothing
Assume that we are currently in period t. We calculated the
forecast for the last period (Ft-1) and we know the actual
demand last period (At-1)
Ft = Ft-1 + α ( At-1 – Ft-1 )
The smoothing constant α expresses how much our forecast will
react to observed differences.
Low α : small reaction to differences.
High α : strong reaction to differences.
forcasting methode; Linair regression
- Fitting a straight line to data
- Explaining the change in one variable through changes in other
variables.
Y = aX + b
By using linear regression, we are trying to explore which
independent variables affect the dependent variable
Y: Dependent variable (=output)
X: Independent variable (=input)
comparing forcasting methodes
Mean Absolute Deviation (MAD)
forecast error = Difference between actual and forecasted value
−
