NANA (perceptron neural net prototype). Developed as my diploma in 2008-2009, this project is research and implementation of artificial intelligence (neural nets), math modelling and data science.
I used as data source global sealevel data for a last 100 years (where it possible) to learn neural net claster to make a predictions for short and long period of time depend of input (geodata).
Program allows to analyze data and neural net structure (for different type of learning curves, where in general used back propogation algorythm) and improve predictions and/or reconfigure network structure itself.
Program written on Borland C++ Builder 6 and uses BDE database engine to store data.
