from dataset.store import loadCSV import numpy as np from sknn.mlp import Classifier, Layer from sklearn.cross_validation import train_test_split from sklearn.metrics import accuracy_score from libs.roc import getScore from store import saveXLSX from sklearn import metrics from datetime import datetime folder = 'dataset/2 classes/breast-cancer-wisconsin' file = 'breast-cancer-wisconsin.csv' # ==================================================== data = np.asarray(loadCSV(folder, file)) y, X = data[:, 1], data[:, 2:].astype(float) y[y == 'M'] = 0 y[y == 'B'] = 1 y = y.astype(int) # ==================================================== learning_rate = [0.01] # [0.001, 0.005, 0.01, 0.05] learning_rule = ['sgd'] # ['sgd', 'adagrad'] hidden_units = [8] # [8, 16, 32, 64, 128, 256, 512] n_iters = [16] # [16, 32, 64, 128, 256, 512, 1024] output = {} output['score'] = [[
from dataset.store import loadCSV import numpy as np from sknn.mlp import Classifier, Layer from sklearn.cross_validation import train_test_split from sklearn.metrics import accuracy_score, confusion_matrix from libs.roc import getScore from store import saveXLSX import datetime folder = 'dataset/3 classes/wine' file = 'wine.data.csv' # ==================================================== data = np.asarray(loadCSV(folder, file)) y, X = data[:, 0].astype(int), data[:, 1:].astype(float) y[y == 1] = 0 y[y == 2] = 1 y[y == 3] = 2 # ==================================================== learning_rate = [0.001, 0.005, 0.01, 0.05] learning_rule = ['sgd', 'adagrad'] hidden_units = [8, 16, 32, 64, 128, 256, 512] n_iters = [16, 32, 64, 128, 256, 512, 1024] output = {} output['score'] = [['learning_rate', 'learning_rule', 'hidden_units', 'n_iters', '#', 'acc', 'u', 'VUS_1', 'VUS_2', 'TP1', 'F12', 'F13', 'F21', 'TP2', 'F23', 'F31', 'F32', 'TP3' 'countingtime']]
import numpy as np from dataset.store import loadCSV, saveXLSX from sklearn.cross_validation import train_test_split from sklearn.metrics import accuracy_score, confusion_matrix import tensorflow as tf import math from datetime import datetime from libs.roc import getScore from sklearn import metrics folder = 'dataset/2 classes/german_credit' file = 'german_credit.csv' # ==================================================== data = np.asarray(loadCSV(folder + '/' + file)) y, X = data[:, 0], data[:, 1:].astype(float) y = np.asarray([[1, 0] if (_ == '0') else [0, 1] for _ in y]) # ==================================================== learning_rate = [0.001] # [0.001, 0.005, 0.01, 0.05] learning_rule = ['adagrad'] # ['adagrad', 'sgd'] hidden_units = [8] # [8, 16, 32, 64, 128, 256, 512] hidden_layers = [2] # [2, 3, 4] n_iters = [16] # [16, 32, 64, 128, 256, 512, 1024] # ==================================================== def initWeight(size): return tf.Variable(tf.zeros(size)) def initWeight_U(size):
import numpy as np from dataset.store import loadCSV, saveXLSX from sklearn.cross_validation import train_test_split import tensorflow as tf import math from datetime import datetime from libs.roc import getScore from sklearn import metrics folder = 'dataset/2 classes/breast-cancer-wisconsin' file = 'breast-cancer-wisconsin.csv' data = np.asarray(loadCSV(folder + '/' + file)) y, X = data[:, 1], data[:, 2:].astype(float) y = np.asarray([[1, 0] if (_ == 'M') else [0, 1] for _ in y]) # ==================================================== learning_rate = [0.001, 0.005, 0.01, 0.05] learning_rule = ['sgd', 'adagrad'] hidden_units = [8, 16, 32, 64, 128, 256, 512] n_iters = [16, 32, 64, 128, 256, 512, 1024] # ==================================================== def initWeight(size): return tf.Variable(tf.zeros(size)) def initWeight_U(size): minR, maxR = 0, 0 if (len(size) == 1): minR, maxR = -1/size[0], 1/size[0] elif (len(size) == 2):