What Is Validation?
Validation is a must for any science (Strict Primal-Dual Approach). Currently, the process to validate complex data analysis is restricted to testing uncertain results against uncertain database information, which is an imprecise method.
True validation is using a alternate method (Dual) to prove the calculation results (Primal). The multiplication function (e.g. 2×3=5) can be validated via the addition function (2+2+2≠5). In another example, the value of sinθ can be validated by calculating cosθ and adding their squares: sin²θ+cos²θ=1, where the “1” must be an integer(!).
This type of Primal-Dual Validation can be extended to all calculations and achieved in any complex data analysis. Unfortunately, whether statistics, machine learning or another modern data analysis process, true validation procedures are not standard and might not even be viable.
In OI, the absolute numerical validation of every step in the combinatorial procedure is mandatory to guarantee Numerically Certain results.
Why Should You Care?
The lack of comprehensive validation leads to uncertain results. Uncertain results waste your bandwidth and resources by giving different results from the same data set, leading to false positives, false negatives, and a constant state of trying to track down the source of elusive problems. Consequently, one cannot trust or rely on invalidated data analysis results.
Driving in an autonomous vehicle that uses invalidated numerical operations is very risky and makes the root cause of issues indeterminate. Using invalidated data analysis methods to diagnose illness or create medical solutions is foundationally arbitrary.
Because OI is built on a foundation of absolute Numerical Validation (Strict Primal-Dual Approach) you can have absolute trust in the reliability of the results it provides.
How Does Inora Use Validation To Help You?
Validation ensures the numerical certainty in the results of any output calculated by OI Core Technology. All software products licensed by Inora utilize validation. That means you don’t have to question the results or values you get and you can focus on actually improving root cause issues instead of having to second-guess results and wonder where issues are coming from.