Measuring (or estimating) the correlation between only two variables and exploiting it to take the optimal decision is simple for someone with training in statistics and decision theory.
But it is much more difficult when you have many variables, all potentially correlated with one another. For example, if you have 100 variables, you have to estimate simultaneously 4950 correlations between 4950 different pairs of variables. Very few people have enough data to estimate accurately such a large number of correlations. In practice, trouble often starts when you have more than 30 correlated variables.
Errors in estimating correlation coefficients from available data are not just randomly distributed. They conspire to make you systematically take the wrong decision. This phenomenon is known as error maximization. It can be fatal, unless the appropriate remedies are put in place.
Over the years, Studdridge International has developed proprietary technology to accurately estimate large numbers of correlation coefficients, and to make optimal decisions based on these correlations. We are world leaders in detecting and solving the problem of error-maximization.
Fields of application include wireless communication, finance, radars, biotechnology, etc.
We partner with you, our clients, to deliver custom-made solutions that exploit large data sets optimally.
We make all the variables in your database sing in unison.