A Multi Layer Perceptron created to calculate XOR (eXclusive OR) on 5 inputs.
- Bad facts are those where the actual output calculated by the neural netowrk is different from the expected output.
- Good facts - negation of bad facts i.e. actual = expected
Refer to this link for more information about neural networks and MLPs.
- 5XOR.csv - the XOR entries/rows/combinations for 5 inputs with their outputs
assign neural network.py
is the only program.BadFactsVsEpochResults.csv
is an example of the results exported to a spreadsheet.- The rest are many graphs which are images.
- Graph data is affected by how the program is written, for example if the order of XOR entries is randomised. In fact, there are 2 folders, one named graphs before randomising trainingSet and graphs.