#Create directory if directory does not exist
filepath = '../../figs/out/%s/%s/%s/' % (scriptname, nowdate, dataset)
if not os.path.exists(filepath):
os.makedirs(filepath)
# Import dataset and target
if dataset == 'MESA':
X = pd.read_csv(
'../../data/mesa/MESA_Clinical_data_full_COMBI-BIO_non-verbose.csv',
sep=',',
header=0,
index_col=1)
X = p2.filt_imp(X, 0.1)
else:
X = pd.read_csv('../../data/simulated/mvnsim/mvnsim' + dataset +
'.csv',
sep=',',
header=0,
index_col=0)
y = np.load('../../data/simulated/mvnsim/target' + dataset + '.npy')
#print(y)
#print(y.shape)
#print(X.shape)
'''
p2.distribution_boxplot(X,
y,
scriptname = 'mesa_kpca2'
#List of datasets to test
#dataset_list = ['diabetes', 'sex', 'cac_binomial', 'cac_extremes', 'family_hx_diabetes', 'parent_cvd_65_hx', 'family_hx_cvd', 'bp_treatment', 'diabetes_treatment', 'lipids_treatment', 'mi_stroke_hx', 'plaque']
dataset_list = ['diabetes', 'sex', 'cac_binomial']
for dataset in dataset_list:
print('\n##### Now running dataset %s #####' % dataset)
#Create directory if directory does not exist
filepath = '../../figs/out/%s/%s/%s/' % (scriptname, nowdate, dataset)
if not os.path.exists(filepath):
os.makedirs(filepath)
X = pd.read_csv(
'../../data/mesa/MESA_CPMG_MBINV2_ManuallyBinnedData_BatchCorrected_LogTransformed_1stcol_%s.csv'
% dataset,
sep=',',
header=None,
index_col=0)
#print(X)
X_imp = p2f.filt_imp(X, 0.1)
X_imp_df = pd.DataFrame.from_records(X_imp)
#print(X_imp_df)
X, y = p2f.tsplit(X_imp_df)
#print(y)
X_scaled = scale(X)
#Create directory if directory does not exist
filepath = '../../figs/out/%s/%s/%s/' % (scriptname, nowdate, dataset)
if not os.path.exists(filepath):
os.makedirs(filepath)
# Import dataset and target
if dataset == 'MESA':
X = pd.read_csv(
'../../data/mesa/MESA_Clinical_data_full_COMBI-BIO_non-verbose.csv',
sep=',',
header=0,
index_col=1)
X = filt_imp(X, 0.1)
else:
X = pd.read_csv('../../data/simulated/mvnsim/mvnsim' + dataset +
'.csv',
sep=',',
header=0,
index_col=0)
y = np.load('../../data/simulated/mvnsim/target' + dataset + '.npy')
#print(y)
#print(y.shape)
#print(X.shape)
distribution_boxplot(
X,
# Collect optimal tier1 gammas
opt_t1_gammas = []
#Using first input dataset to generate toy datasets
inp_df = pd.read_csv(
'../../data/mesa/MESA_CPMG_MBINV2_ManuallyBinnedData_BatchCorrected_LogTransformed_1stcol_%s.csv'
% inp_dataset_list[0][1],
sep=',',
header=None,
index_col=0)
print(
'\nUsing %s dataset to generate simulated datasets for the purpose of tuning algorithms and hyperperameters.'
% inp_dataset_list[0][0])
X_imp = p2f.filt_imp(inp_df, 0.1)
X, y = p2f.tsplit(X_imp)
toy_dataset_list, toy_y = p2f.toybox_gen(X)
for toy_label, toy_X in toy_dataset_list:
print('\n##### Now running dataset %s through tier 1 #####' % toy_label)
#Create directory if directory does not exist
filepath = '../../figs/out/%s/%s/%s/' % (scriptname, nowdate, toy_label)
plotpath = '%splotting/' % filepath
if not os.path.exists(filepath):
os.makedirs(filepath)
os.makedirs(plotpath)
import pandas as pd
import numpy as np
from numpy.random import multivariate_normal
import p2funcs as p2f
inp_df = pd.read_csv(
'../../data/mesa/MESA_CPMG_MBINV2_ManuallyBinnedData_BatchCorrected_LogTransformed_Data.csv',
sep=',')
imp_df = p2f.filt_imp(inp_df, 0.1)
'''
def toybox_gen(inp_df):
#Define size of component using input data
cols, rows = inp_df.shape
comp_size = [cols, int(round(rows/2))]
cov1 = np.array([[[0.3, 0.2], [0.2, 0.2]], [[0.6, 0.4], [0.4, 0.4]], [[1.2, 0.8], [0.8, 0.8]], [[2.4, 1.6], [1.6, 1.6]], [[6, 4], [4, 4]],[[9, 6], [6, 6]]])
cov2 = np.array([[12, 8],[8, 8]])
mean1 = np.array([[20, 15], [20, 15], [20, 15], [20, 15], [20, 15], [20.5, 15.5]])
mean2 = [20, 15]
#Set up target array
target = np.zeros(100, dtype=int)
target[0:50] = '1'
dataset_list = []
# counter for labelling dataset
counter = 1
