02. Forecasting (Unfinished)

Question 1

why do we need forecasting?

Answer 1

• Resource planning: e.g. Staffing
• Financial management: Revenue / yield cash flow
• Inventory management: food & beverage, housekeeping supplies
• Capital management

Question 2

what is time series

Answer 2

Define: a set of variables measured at a continuous point in time. A set of values of a variable measured over time
e.g.
- number of customers per day in a hotel
- Daily demand of croissants in a bakery
- Number of visitors a day in Musée du Louvre
Used to forecast:
- How many customers will come to my restaurant tomorrow?
- How many burgers can I sell tomorrow?

Question 3

4 Major components of Time series

Answer 3

1. Trends: the gradual increase or decrease of a time series over a longer period of time
2. Seasonality: regular pattern of variability within certain time periods, Such as year/week (Olympic is not seasonal! Happens Once every 4 years)
3. Cycle: any regular pattern above and below the trend line lasting more than one year is part of cyclical component. It’s basically a longer seasonality
4. Random Variations: caused by black swan events, in unusual situations.

Question 4

What is human distortions

Answer 4

Human tendancy to have different approaches to forecasting. Optimitstic, Neutural, Pessimistic way to forecast

Question 5

What are the two forecasting approaches?

Answer 5

Subjective (affected by human)

Mathematical Approachs

Question 6

What is Extrapolative methods?

Answer 6

it seeks to identify the patterns that lies within the past data. Patterns depended on our 4 components of time series. Hence allowing us to make predictions in the future.

• A subjective forecasting method could be affected by human distortions
• A mathematical forecasting method is based on a past time series
• A mathematical forecasting method eliminates human distortions

Question 7

What are the two families of the MATHMETICAL methods?

And which one is for shrot term which one is for long term?

Answer 7

Smoothing methods
- Simple, reduce random elements, used for short-term forecasting.
- It could be applicable for short-term time series.
Decomposition methods
- Complex
- Also eliminates random elements (outliners)
- Used for long-term forecasting
- Isolating trend & seasonality / cycle

Question 8

WHat are the three forecasting methods we will learn in this semester?

Answer 8

Three methods presented in this semester:

• Simple moving average (SMA)
• Weighted moving average (WMA)
• Simple exponential smoothing (SES)

Question 9

what are the 3 Notions used for this course?

Xt; Ft; Wt

Answer 9

X_t: Historical data used in time series, e.g. X_1: Observation for period 1
F_t: Forecasting for period t.
W_t: Weight for the observation of period t.

Question 10

For mular for SMA

Answer 10

all the ovservations added together and divided by the number of observations

Question 11

formular of the weighted moving average

Answer 11

F4= W1*X3 + W2 *X2 + W3 * X1

Question 12

Formular for SES

Answer 12

alpha * most recent observations + (1-alpha) * most recent forecasted value

Question 13

what are the two methods the decomposition method includes ?

Answer 13

Linear regression

and Seasonal Index

Question 14

What is the equation for Linear Regression?

Answer 14

f(x) = ax + b