For one of my Artificial Intelligence projects, we had to study neural networks and how adjusting their initial properties effect their outcome.
To test this, we used two separate data sets, the age of Abalone’s (https://archive.ics.uci.edu/ml/datasets/abalone) and the the classification of Iris flowers (https://archive.ics.uci.edu/ml/datasets/iris)
Unfortunately, for this project we did not have time to use our own Neural Network. Instead we used WEKA, developed by the University of Waikato.
My findings and full paper can be found in the below linked pdf, and all of my datasheets can be downloaded at my github here: https://github.com/QuantumInfinite/Feed-forward-neural-networks-for-data-classification