TIME SERIES METHOD p 65 - 70
TIME SERIES
- Time ordered sequence of observations taken at regular intervals of time
• Hourly daily weekly monthly
• Forecsting based on time series data are made on the assumpt
that future values of the series can be estimated from their ow
past values
• Requires identifying underlying behaviour of the series
® Plotting data and visually examining it
® Can be random or irregular one time variations
PATTERNS IN TIME SERIES
1. Level
- Average
- Or constant, refers to horizontal pattern of time series
2. Trend
- Steady upward or downward movement
} Population growth, increasing income, cultura
change account for such movement
3. Seasonality
- Regular variations related to the time of year or day
- Mostly linked to weather or recurring events
} Ice cream = higher sales in the summer
} Restos, supermarkets and theates eperience
weekly and even daily seasonal variations
4. Cycles
- Wavelike variations lasting more than one year
} Economic, polititcal, agricultural conditions
} Ex) supply of cattle
- Upward or downward pattern in demand that repeats in
,on
, } Restos, supermarkets and theates eperience
weekly and even daily seasonal variations
4. Cycles
- Wavelike variations lasting more than one year
} Economic, polititcal, agricultural conditions
} Ex) supply of cattle
- Upward or downward pattern in demand that repeats in
longer intervals
- Also related to policy, product life
} Ex) sales of sports gear during World Cup
} New construction projects near elections
5. Irregular variation
- Caused by unusual circumstances, not reflective of typica
behaviour
} Severe weather conditions, strikes, sales
promotions
} Whenever possible should be identifies and
removed from the data
6. Random variation
- Residual variations after all other behaviours are account
for
- Called noise
- Randomness comes from combined influence of relativel
unimportant factors
- Cannot be reliably predicted
- Time series techniques smooth out random variations in
data
HOW TO CALCULATE
NAÏVE METHOD
- Next period forecast = last period demand
- Simple to use and understand
TIME SERIES
- Time ordered sequence of observations taken at regular intervals of time
• Hourly daily weekly monthly
• Forecsting based on time series data are made on the assumpt
that future values of the series can be estimated from their ow
past values
• Requires identifying underlying behaviour of the series
® Plotting data and visually examining it
® Can be random or irregular one time variations
PATTERNS IN TIME SERIES
1. Level
- Average
- Or constant, refers to horizontal pattern of time series
2. Trend
- Steady upward or downward movement
} Population growth, increasing income, cultura
change account for such movement
3. Seasonality
- Regular variations related to the time of year or day
- Mostly linked to weather or recurring events
} Ice cream = higher sales in the summer
} Restos, supermarkets and theates eperience
weekly and even daily seasonal variations
4. Cycles
- Wavelike variations lasting more than one year
} Economic, polititcal, agricultural conditions
} Ex) supply of cattle
- Upward or downward pattern in demand that repeats in
,on
, } Restos, supermarkets and theates eperience
weekly and even daily seasonal variations
4. Cycles
- Wavelike variations lasting more than one year
} Economic, polititcal, agricultural conditions
} Ex) supply of cattle
- Upward or downward pattern in demand that repeats in
longer intervals
- Also related to policy, product life
} Ex) sales of sports gear during World Cup
} New construction projects near elections
5. Irregular variation
- Caused by unusual circumstances, not reflective of typica
behaviour
} Severe weather conditions, strikes, sales
promotions
} Whenever possible should be identifies and
removed from the data
6. Random variation
- Residual variations after all other behaviours are account
for
- Called noise
- Randomness comes from combined influence of relativel
unimportant factors
- Cannot be reliably predicted
- Time series techniques smooth out random variations in
data
HOW TO CALCULATE
NAÏVE METHOD
- Next period forecast = last period demand
- Simple to use and understand
