Business Forecasting Topic 4 Flashcards

1
Q

Global models of time series

A

assume same trend or seasonal pattern applies to entire time series

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

local models of time series

A

assume different trends or seasonal patterns can apply at different points in time

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

exponential smoothing models

A

allow forecasts to take into account any changing patterns by updating estimates of patterns as new observation is available

  • older/later data = less weight
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Simple (or single) exponential smoothing

A
  • where series has flat trend
  • subject to changes to new levels
  • estimate current underlying level of time series (level when noise filtered out)
  • estimate of current lvl = forecast for next period
  • smooth out random fluctuations in series (not eliminate randomness)
  • only works if not a trended series otherwise all same values
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

SES forecasts formulae

A

ON SHEET!!
forecast = weighted average of latest observation and current forecast

  • larger α = greater weight attached to most recent observation

more than 1 period ahead = one period ahead forecasts

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

initial forecast for SES

A

required to start process
- average of past observations
- educated guess

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

smoothing constant

A

α
0 = smooths thoroughly
1 = naive forecast

  • value chosen to balance stability (not overact to freak figures) and sensitivity (respond to new conditions)
  • α value with lowest MSE = produce future forecasts
  • MSE over the fitting period, find optimal α to minimise MSE
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

high α value

A

α = 1 = naive forecasts = shift 1 period
behave exactly like data
respond to sensitivity

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

low α value

A

dont catch up data

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Exponentially weighted moving averages (EWMA)

A

each forecast = weighted average of all previous observations

most recent = highest weight
oldest = lower weight

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

forecasting more than 1 period ahead

A

SES assume no upward or downward trend in time series = extrapolate flat line in future

  • forecasts of more than 1 period ahead are identical to the one period ahead forecast
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

advantages of SES

A
  1. simple
  2. easily automated
  3. low data storage req
  4. more accurate than many complicated methods
  • large number forecast on regular basis
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

tracking signals

A
  • detect when forecasts = inaccurate
  • α value needs to be adjusted

error high = tracking signal high = showing a widening gap between data and forecasts

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

cumulative error

A
  • if growing -> ideally values close to zero if unbiased forecasts
  • alerts us when things are wrong
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

issues with using cumulative error as method of alerting of issue

A
  1. how large/small (small = -ve) does error have to be before alert
  2. one large/unrepresentative error keeps it high/low -> even if subsequent forecasts are unbiased
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Smoothed error tracking signal formulae

A
  • overcomes 2 problems of cumulative error

equation 1 = smoothes error by accounting for sign, expect close to 0 (errors of unbiased forecasts cancel each other)

tracking signal = close to zero = assume no problem

17
Q

Mt

A

estimate of mean absolute errors

forecast error large = mean absolute error = large = tolerant of smoothed mean error straying quite way from 0 before think there is problem

18
Q

adaptive response rate exponential smoothing (ARRSES)

A

tracking signal large = forecast failed to adapt, more sensitive = increase α
tracking signal small = coping well - small α will suffice

value of α varies denote value at period t

19
Q

formulae for ARRSES

A
20
Q

ARRSES strengths and limitations

A
  • completely automated
  • more attractive than SES if large numbers on regular basis
  • unstable
21
Q

one period ahead forecast

A

main reason for = take into account the latest data

22
Q

data has continuous upward or continuous downward trend

A

SES forecasts will never be able to catch up with data

  • forecasts cant catch as data is faster (increasing/decreasing)
  • highest possible forecast, upward trend, α = 1 = naive forecast upward trend = forecast lagging behind data by 1 period