Gene Expression Programming
Learning project to explore both the construction of this GP family machine learning algorithm and to grok the extent of F#'s abilities to express a domain concisely.
The ultimate aim of the project is to produce a viable program that can be used to generate neural networks or other functions to explore relationships in data.
Current stage is constructing the first nuts and bolts:
So far it has some data structures, rudementary translation of chars to symbols with index & arity ( a record type), a basic algorithm for chopping a multigenic chromosone into chosen head length and the beginnings of the chromosone processor.
Next stage will be to implement a gene to expression tree parser followed by the multigenic ET parser.
Once complete, further work such as mutation will be explored.