Using Diveplane Reactor, businesses can intelligently automate repetitive tasks in order to allow analysts to spend more time on the complex, harder-to-solve problems that require a human’s intervention. The output of the Reactor platform is comprehensive, defensible, and clear about how it arrived at a certain decision and exactly what data informed that choice. Through human review, Reactor allows an organization to pinpoint and analyze potentially highly biased data and remove it from all future analysis. This experience is entirely different from the black box systems that currently dominate the market. Reactor removes the black box entirely and replaces it with something better altogether.
The explanations that Reactor provides to explain its decisions come from a series of proprietary measurements such as our patented “conviction” metric, which represents how surprised the system is by new data. A higher conviction means the system is fairly convinced that new data is unsurprising and most likely fits the pattern of the other data, whereas a low conviction means the model is very surprised to see the new data and believes it is too different to belong in the dataset.