], '<DIGIT>' : [ {'become' : '<DIGIT><DIGIT>', 'p' : 1}, {'become' : '0', 'p' : 1}, {'become' : '1', 'p' : 1}, {'become' : '2', 'p' : 1}, {'become' : '3', 'p' : 1}, {'become' : '4', 'p' : 1}, {'become' : '5', 'p' : 1}, {'become' : '6', 'p' : 1}, {'become' : '7', 'p' : 1}, {'become' : '8', 'p' : 1}, {'become' : '9', 'p' : 1} ] } ge.startExpr = '<EXPR>' trainingSet = { 'header' : ['otsu', 'stdev', 'mean', 't', 'minOtsu'], 'target' : 't', 'data' : thresholdingData } ge.trainingSet = trainingSet ge.train() ge.printAllPhenotype() #good features should have correlation with the output (done), #good features should not be correlated each other