def minimize_linregr(expected_xs_dict, err_func, stall_precision, max_stalled_in_row, ga_options,
return_classify = False, test_classification = False):
error_f = lambda coefs: sum([ err_func(linregr(coefs, xs), expected)
for expected, xss in expected_xs_dict.iteritems()
for xs in xss
])
res = ga.minimize_until_stall(error_f, stall_precision, max_stalled_in_row, ga_options)
res_ = (res, )
if return_classify or test_classification:
classify = linregr_classifier(expected_xs_dict.keys(), res['best'][0])
res_ = res_ + (classify, )
if test_classification:
test_results = [ (clazz, [classify(xs) == clazz for xs in xss])
for clazz, xss in expected_xs_dict.iteritems()
]
test_count = map(lambda (clazz,rs): (clazz, count(identity, rs), len(rs)), test_results)
test_results = map(lambda (clazz,c,n): (clazz, str(c) + " of " + str(n)),
test_count
)
test_results_total = reduce(lambda (ac,an), (_,c,n): (ac+c, an+n), test_count, (0,0))
test_results.sort(key=lambda(c,_): c)
res_ = res_ + (test_results, test_results_total)
return res_
def ga_test_s_2():
_print_2()
res = ga.minimize_until_stall(_fit_f_2, 1e-5, 40, _opts_2)
print "Result:"
print res
def ga_test_s_1():
_print_1()
res = ga.minimize_until_stall(_fit_f_1, 1e-5, 40, _opts_1) # fitness_func, stall_precision, options, max_iter = None)
print "Result:"
print res