Next Generation AI in Data Analysis
Conventional linear correlation measures, often used in association with designing a portfolio and managing risk, may provide misleading results.
Optimum adopts a more relevant measure of correlation - generalized correlation - which accounts for non-linear aspects of data. The method is based on proprietary model-free cognitive AI technology which transforms data scatter plots into images, emulating an expert actually looking at data. This is AI 2.0.
The huge advantage of generalized correlations are numerous. First of all, they capture correlations that classical techniques can miss. Secondly, they discard correlations that are identified mistakenly by conventional means. Finally, they provide realistic values of correlation unlike, for example, Pearson correlations, which can often be excessively high.
However, the biggest advantage of generalized correlations is that they don’t require a cut-off value in order to differentiate between significant and negligible values thereof. The method of computation naturally establishes which correlations are significant and which are not.
The method for the computation of generalized correlations has its roots in quantum physics and nonlinear mechanics. For more than a decade, our technology is being used in numerous sectors, such as defense, manufacturing and medicine.
What We Do With Data
Data is analyzed by emulating the brain as it ‘sees it’ in the form of an image, without the need to build math models. This allows us to determine the true structure and nature of data.
Our measure of correlation enables us to better reflect the level of risk associated with an individual company, portfolio of companies or other financial instruments or asset classes.