We developed a neuro-bio inspired generic deep learning algorithm that works as simple as an intelligence test. It automatically detects patterns in time series and allows you to make predictions.
While other ML approaches become more and more complex we wanted to build a model that is straightforward and that allows you to see where predictions are coming from. That is why our algorithm is not based on any existing frameworks. Instead we started from scratch and defined a model consisting of neurophysiological mechanisms and elements like neurons, axons, dendrites. We combine those elements to a neuronal network that allows information to be represented and predictions to be made.
While our implementation provides an api interface it is entirely written in C#, does not require a database or other services and it can even run without internet connection.
The model supports temporal as well as spacial patterns. That means you can provide a context within your time series.
The model can be adapted to your needs.
As the algorithm can be adapted to almost any industry we would like to give you some examples of what kind of applications can be realized with our model.