Read me

Sourcecode for the manuscript 'Prediction of Suitable Catalysts for the OCM Reaction by Combining an Evolutionary Approach and Machine Learning'

To change the generationsize: Change P To change the number of generations: Change G To change the evolutionary factor: Change E

To choose if the best, the worst or a random picked original parent generation is used, change the string parent_selection_mode to 'best','worst' or 'random'

Fast generates a lot of directed evolutions all with the same parameters without the figures.

EvaluationCatalysts.ipynb evaluates the results of mutations generated by either the unimproved or the improved random forest. Each mutation has to be saved according to the following file format:

folder_name = 'random_'+str(g) # or 'best', 'worst' file_name_csv = 'Mutants_19_'+str(g)+'.csv' #Change number to G-1 according to Fast

All Codes should be started in their own directory, because they generate folders with the same names. Data will be overwritten or mixed.

Most graphs are produced as pgf files. 'lualatex' might be a better option for some people.