 |
Download a pdf copy of this help file here |
|
Download a pdf copy of this help file here |
VoseTSEmpiricalFit({data},Multiply,LastValue,Uncertainty)
Array function that returns random values from an empirical time series directly constructed from data. You can choose between assuming an additive or a multiplicative random walk with the Multiply parameter.
• {data} - array with historical data to construct the time series from.
• Multiply - optional boolean parameter (TRUE/FALSE). Set TRUE to assume a multiplicative random walk and FALSE (default) for an additive random walk.
• LastValue - (optional) last known historical value. The generated random values will continue from this value onwards. If no LastValue parameter is specified, the last element of the {data} array is taken as last known historical value.
• Uncertainty - optional boolean parameter (TRUE/FALSE). Set to TRUE to incorporate the uncertainty that always exists about the constructed time series, and to FALSE (default) to use a fixed "best guess" empirical time series.
The time series is constructed by looking at the increments between each historic data value and its immediate past value. When Multiplicative=TRUE the increments are ratios, when FALSE they are differences.
A non-parametric continuous distribution is then constructed on the observed increments: first-order (Ogive) if Uncertainty=FALSE and second-order (OgiveU) if Uncertainty=TRUE.
We then use randomly sampled values from this increment distribution to generate the output , starting onwards from LastValue.
Example
When Multiply is set to TRUE, the first value of the output array is:
LastValue*Increment
where Increment is a random sample from the constructed increments distribution.
When Multiply is set to FALSE, the first value of the output array is
LastValue+Increment
where Increment is a random sample from the constructed increments distribution.