import os
import sklearn.model_selection as ms
from helpers import get_data_from_csv
RANDOM_STATE = 42
for d in ['BASE', 'RP', 'PCA', 'ICA', 'RF']:
n = './{}/'.format(d)
if not os.path.exists(n):
os.makedirs(n)
OUT = './BASE/'
# import wine quality data
wineX, wineY = get_data_from_csv("winequality-white.csv",
n_features=11,
sep=';') # instances
real_Y = []
for Y in wineY:
for y in Y:
real_Y.append(y)
wineY = np.array(real_Y)
# import semeion data
digitX, digitY = get_data_from_csv("semeion.data.csv",
n_features=256,
sep=' ',
header=None)
real_Y = []
for Y in digitY:
for i in range(10):
import numpy as np
from sklearn.preprocessing import StandardScaler
from sklearn.pipeline import Pipeline
from helpers import nn_arch, nn_reg
from sklearn.neural_network import MLPClassifier
from sklearn.model_selection import GridSearchCV
from sklearn.decomposition import FastICA, PCA
from helpers import get_data_from_csv
out = './ICA/'
np.random.seed(42)
# import wine quality data
wineX, wineY = get_data_from_csv("./BASE/wine_trg.csv",
n_features=11,
sep=',',
header=None)
digitX, digitY = get_data_from_csv("./BASE/digit_trg.csv",
n_features=256,
sep=',',
header=None)
wineX = StandardScaler().fit_transform(wineX)
digitX = StandardScaler().fit_transform(digitX)
clusters = [2, 5, 10, 15, 20, 25, 30, 35, 40]
dims = [2, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60]
dims_wine = [i for i in range(2, 12)]
# data for 1
mpl.use('TkAgg')
import matplotlib.pyplot as plt
import matplotlib.cm as cm
np.random.seed(42)
rds = ['BASE', 'PCA', 'ICA', 'RF', 'RP']
rds = [sys.argv[1]]
for rd in rds:
out = './{}/'.format(rd)
# import wine quality data
if rd == "BASE":
wineX, wineY = get_data_from_csv(out + "wine_trg.csv",
n_features=11,
sep=',',
header=None)
digitX, digitY = get_data_from_csv(out + "digit_trg.csv",
n_features=256,
sep=',',
header=None)
else:
wineX, wineY = get_data_from_csv(out + "wine_datasets.csv",
sep=',',
header=None)
digitX, digitY = get_data_from_csv(out + "digit_datasets.csv",
sep=',',
header=None)
wineX = StandardScaler().fit_transform(wineX)
digitX = StandardScaler().fit_transform(digitX)