Descriptions of algorithms that are able to select time-indices that show certain emergent features. They are the logical bridge into the future because they are popperian universalities in the sense that they are constructed to work in the past and in the future in exactly the same way.
As an example, we choose the loan portfolio of the peer-to-peer lending company Lending Club (122,444 loans).
We can identify subsets of loans, for which multiple emergent laws always were found. For example the loans to A rated customers, which used the borrowed money for repaying existing credit card debt, always had the following properties:
In every sequence – i.e. in every prediction horizon – of TP=15,000 loans:
- the return of the loans was always between 6.9% and 9.6%.
- the return of the loans was always 0.4%-3.3% less than the average return of all loans.
- the return of the loans was always 0.2%-2.2% less than the average return of all A-rated loans.
- the probability of default (PD) was always between 2.9% and 8%.
- and the loss given default (LGD) was always between 22.5% and 39.6%.
Following the simple induction principle:
“Only what has been true until now, can be true in general”
we can make a prediction about many properties of this subset of loans.
Patterns observed in sequences of measurements, that were always followed by a couple of different patterns, are called “Objects” and we can say casually, that the above laws are “attached” to the object.