This is a profiling and stratification tool which uses Nolanlean algebra and peer relativity reshaping to construct an enhanced dataset. Like variables are placed together in a grouping, where a single value is going to represent a row of observations. Firstly for each variable the highest and lowest values are allocated 53 and 1, then all remaining observations are transformed to their relative position. Secondary the reshaped values are sequenced together using the Nolanlean algebra to generate a digital fingerprint hash value which represents the row value for the group. This works as a compression processes. Then the process is repeated as many times as required.
An example would be the values in an taxation form. All the variables would be placed in a grouping of
- Income = (wages + investments + pensions + super)
- Deduction = (travel + education + memberships)
- Offsets = (investments + fuel + negative gearing + rebates)
- Payments = (medicare + landtax + GST + state taxes)
- Other = (excise + rent + childcare + salary sacrificing)
- Demographics = (sex + age + residential location + off shore + marriage status + directorship)
Each tax payer would then receive a stratification index number, which puts them within similar groupings with other tax payers.
