My forecasting method is as much outside the box as my approach to data and images. It is characterised by:
- independence of scale and application, i.e. any data covering any time interval can be used
- the interval of the input data determines the interval of the output
- N input points produce N-1 output points.
The method does not ‘know’ whether the data is regular or irregular and thus does not try to fit a curve. However, since it does forecast regular data with 100% accuracy, the question is how well does it do for irregular and thus unpredictable data, such as financial data and data relating to climate change.
A factorised sine function of 200 points forecast
Yellow input data points produce blue forecasts.
For academics, who want to know how I’m doing it:
- since nobody paid me to develop my method, I want to turn it into money, rather than publish the knowhow.
For investors, who want to ask whether my IP is protected:
- I withdrew five patent applications, after an American IP lawyer advised me to protect myself with trade secrets, for patents are the game of the big boys.
For ‘strategic co-developers’, who want to create an expert portal that addresses to help you solve your problems:
- send me sample data and let’s talk
- the method can ‘learn’ from comparing past forecasts with past realities.
Next: Practical Applications.