def __init__(self,block,stack): # evaluate function inputs inputs = evaluate_inputs(block['inputs'], stack) # run function self.clf = KernelRidge(**inputs['kwargs']) self.clf.fit(*inputs['args'])
def __init__(self,block,stack): # evaluate function inputs inputs = evaluate_inputs(block['inputs'], stack) # run function from sklearn.kernel_ridge import KernelRidge self.clf = KernelRidge(**inputs['kwargs']) self.clf.fit(*inputs['args'])
def LeaveOneOut(block, stack): # evaluate function inputs inputs = evaluate_inputs(block['inputs'], stack) # run function from sklearn.model_selection import LeaveOneOut loo = LeaveOneOut(**inputs['kwargs']) output_dict = {} fold_gen = loo.split(*inputs['args']) output_dict["api"] = loo output_dict["fold_gen"] = fold_gen
def KFold(block, stack): # evaluate function inputs inputs = evaluate_inputs(block['inputs'], stack) # run function from sklearn.model_selection import KFold kf = KFold(**inputs['kwargs']) output_dict = {} fold_gen = kf.split(*inputs['args']) output_dict["api"] = kf output_dict["fold_gen"] = fold_gen return output_dict
def scorer_regression(block, stack): # evaluate function inputs inputs = evaluate_inputs(block['inputs'], stack) output_dict = {} # run function from sklearn.metrics import make_scorer for i in inputs['args']: if i == 'mean_squared_error' or i == 'mean_absolute_error' or i == 'accuracy_score': score_func_1 = i output_dict['score_' + str(i)] = make_scorer( score_func=score_func_1, *inputs['args'], **inputs['kwargs']) else: sys.exit("Function not incoporated as yet") return output_dict
def train_test_split(block, stack): # evaluate function inputs inputs = evaluate_inputs(block['inputs'], stack) # run function from sklearn.model_selection import train_test_split function_output_ = train_test_split(*inputs['args'], **inputs['kwargs']) n_out = len(function_output_) assert n_out == 2 * len(inputs['args']) # create outputs # names are in this order: train1, test1, train2, test2, train3, test3 output_dict = {} for i in range(n_out): if i % 2 == 0: output_dict["train%i" % (int(i / 2) + 1)] = function_output_[i] else: output_dict["test%i" % (int(i / 2) + 1)] = function_output_[i] return output_dict