Chapter 18 - forecasting

Question 1

what is a time series

Answer 1

a series of figures or values recorded over time

Question 2

two components of a time series

Answer 2

1. TREND = underlying long term movement
2. SEASONAL VARIATIONS = short term fluctuations in recorded values due to different circumstances

Question 3

other components of a time series

Answer 3

1. CYCLICAL VARIATIONS = trade cycle related, booms/slumps
2. RESIDUAL VARIATIONS = caused by something other than trend/seasonal/cyclical

Question 4

methods of finding the trend

Answer 4

1. line of best fit
2. linear regression analysis
3. moving averages

Question 5

how can moving averages be used to find the trend

Answer 5

a moving average is an average of the results of a fixed number of periods

it is common for a moving average to be measured over an even number of time periods

by analysing averages, seasonal variations can be smoothed out

Question 6

normal average vs moving average

Answer 6

average = mid point of the period
moving average = relates to particular time period in time series

Question 7

what is the additive model

Answer 7

method of calculating season variation

assumes that:
time series = trend + seasonal variation

Question 8

how to calculate seasonal variation using additive model

Answer 8

1. identify the trend
2. deduct the trend from the time series data to obtain the seasonal variation
   (rearranging time series = trend + seasonal variation into seasonal variation = time series - trend)
3. calculate average seasonal variation for each quarter

Question 9

difference between additive and multiplicative model

Answer 9

• additive model assumes seasonal variation doesn’t increase over time but this is unlikely
• the multiplicative model is more suitable than the additive model for forecasting when the trend is increasing or decreasing over time
• additive model simply adds absolute and unchanging seasonal variations to the trend figures whereas multiplicative multiplies or decreases trend values by a seasonal variation factor

Question 10

how can we use trend and seasonal variation for forecasting

Answer 10

1. identify trend using moving averages, and calculate the average trend growth per period
2. extrapolate trend figures to cover forecast period using expected average trend growth
3. adjust forecast trends by the applicable average seasonal variation to obtain the actual forecast

Question 11

what is regression analysis

Answer 11

an alternative to line of best fit

LINEAR regression finds the line of best fit using mathematical formulae to identify line closest to data points on the diagram

y = a +bx
a = intercept
x = independent variable

Question 12

types of correlation

Answer 12

positive = high values of one variable relate to high values of another

negative = high values of one variable relate to low values of another

Question 13

correlation co-efficient values meaning

Answer 13

r = + 1 = perfectly positively correlated
r = - 1 = perfectly negatively correlated
r = 0 = uncorrelated

Question 14

what is the coefficient of determination

Answer 14

the coefficient of determination r2 measures the proportion of the total variation in the value of one variable that can be explained by variations in the value of the other variable

Question 15

what is spearmans rank correlation coefficient

Answer 15

sometimes variables are in rank order rather than actual values so spearmans rank can be used

same values as correlation coefficient

Question 16

factors affecting reliability of forecasts

Answer 16

• the further into the future it is = more unreliable
• less data available to base it = less reliable
• often assumes what has happened in past will be reliable for future, but not always the case
• usually involves extrapolation outside given observations but this is inaccurate
• danger of random variations upsetting pattern of trend and seasonal variation

Question 17

reliability of regression analysis forecasts

Answer 17

• regression analysis assumes that a linear relationship (creates a straight line when plotted) exists between the two variables when it may not
• assumes value of one variable, Y, can be predicted or estimated from the value of one other variable X. in reality Y may depend on several other variables, not just X