Quantitative Methods
Qualitative forecast often contain huge amount of subjective judgements, as there are lots of personal opinions and other human factors, accurate and reliability are alway the big concerns. Whereas Quantitative methods are more objective and ‘scientific’ . It always involve the historical data, and by using the mathematical models to process those information to found out the patterns embed in the data.
In quantitative model, there are 3 categories of methods:
- Time series (within syllabus)
- Causal
- Simulation
Simulation is usually rely on computer software to complete the tasks, the benefit is it can simulate the situation from reality into the virtual world, where analyst can see the future change in specify situation by changing the relevant parameters, therefore, it is extremely flexible, and not direct loss although faults happen. However, the softwares are normally very expensive, and require specify skill training to use.
In forecasting the prediction can usually be written as:
Forecast = Pattern(s) + Randomness
When a average pattern is embody the data, some variance might appear between the forecast method applied and the actual situation, this variance is call randomness. Most of the method determine that patterns are mainly constituted by four elements: trend, cyclical, seasonality and randomness. The analysis of patterns over times is called Time Series Analysis, and the equation can be written as :
Forecast = Trend t-1 + Seasonality t-1 + Cyclical t-1 + Randomness; where t = time
(Chase, 2009)
(Chase, 2009)
Time series methods are attempts to predict the future patterns by using historical data which over a period time, and here will first introduce three most basic time-series methods, and follow by examples for better illustration:
Copyright © 2015 | University of Warwick, Coventry | [email protected]
POWERED BY KASPAR CHEN