def main(problem, job):
global trainingInputs, trainingOutputs, testingInputs, testingOutputs
random.seed(1000 + job)
#read training data
# training and testing file
dir = '/home/pta/Dropbox/uci/regression/'
# training and testing file
train = dir + problem + ".training.in"
test = dir + problem + ".testing.in"
lines = open(train).readlines()
trainingInputs = []
trainingOutputs = []
for line in lines[1:]:
xs = line.split()
trainingInputs.append([float(x) for x in xs[:-1]])
trainingOutputs.append(float(xs[-1]))
#N = len(trainingOutputs)
# read testing data
lines = open(test).readlines()
testingInputs = []
testingOutputs = []
for line in lines[1:]:
xs = line.split()
testingInputs.append([float(x) for x in xs[:-1]])
testingOutputs.append(float(xs[-1]))
pop = toolbox.population(n=POPSIZE)
hof = tools.HallOfFame(1)
stats_fit = tools.Statistics(lambda ind: ind.fitness.values)
stats_size = tools.Statistics(len)
mstats = tools.MultiStatistics(fitness=stats_fit, size=stats_size)
mstats.register("avg", numpy.mean)
mstats.register("std", numpy.std)
mstats.register("min", numpy.min)
mstats.register("max", numpy.max)
pop, log, best_fitness_each_gen, num_distinct_ind, avg_semantic_distance = algorithms.eaPTA(pop, toolbox, CROSSOVER_RATE, MUTATION_RATE, NUMGEN, stats=mstats,
halloffame=hof, verbose=False)
trainingError = evalSymbReg(hof[0])[0]
testingError = describe(hof[0])
print 'job ', job, ': fitness=', trainingError, '; fittest=',testingError, '; size=',len(hof[0])
# print 'training error: ', trainingError
# print 'testing error: ', testingError
# print log
return job, trainingError, testingError, len(hof[0]), best_fitness_each_gen, num_distinct_ind, avg_semantic_distance
def gp_run(pop):
hof = tools.HallOfFame(1)
stats_fit = tools.Statistics(lambda ind: ind.fitness.values)
stats_size = tools.Statistics(len)
mstats = tools.MultiStatistics(fitness=stats_fit, size=stats_size)
mstats.register("avg", numpy.mean)
mstats.register("std", numpy.std)
mstats.register("min", numpy.min)
mstats.register("max", numpy.max)
pop, log, best_fitness_each_gen, num_distinct_ind, avg_semantic_distance = algorithms.eaPTA(pop, toolbox, CROSSOVER_RATE, MUTATION_RATE, NUMGEN, stats=mstats,
halloffame=hof, verbose=False)
trainingError = evalSymbReg(hof[0])[0]
testingError = describe(hof[0])
print 'job ', job, ': fitness=', trainingError, '; fittest=',testingError, '; size=',len(hof[0]), '; distinct ind= ', num_distinct_ind
# print 'training error: ', trainingError
# print 'testing error: ', testingError
# print log
return job, trainingError, testingError, len(hof[0]), best_fitness_each_gen, num_distinct_ind, avg_semantic_distance
