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,
y,
"Initial category 1 distribution of dataset %s" % dataset,
"Initial category 0 distribution of distribution %s" % dataset,
#output='show'
#output='plotly',
#ply_title="Initial distribution of dataset %s" % dataset,
output='save',
path='%sinitialdist_%s.png' % (filepath, nowtime))
print('\nBoxplot of initial data for dataset %s saved.' % dataset)
## PREPROCESSING ##
#Scale initial data to centre data
X_scaled = scale(X)
X_scaled_df = pd.DataFrame.from_records(X_scaled)
distribution_boxplot(
#path='%sinitialdist.png' % filepath,
)
print('\nBoxplot of initial data for dataset %s saved.' % dataset)
'''
## PREPROCESSING ##
#Scale initial data to centre data
X_scaled = scale(X)
X_scaled_df = pd.DataFrame.from_records(X_scaled)
distribution_boxplot(X_scaled_df,
y,
"Scaled category 1 distribution of dataset %s" % dataset,
"Scaled category 0 distribution of dataset %s" % dataset,
#output='show'
output='save',
path='../../figs/out/%s/%s/scaledist.png' % (scriptname, dataset)
)
#print(X_scaled.shape)
print('\nBoxplot of scaled data for dataset %s saved.' % dataset)
#Initiate KPCAwith various kernels
# As I'm using 500 variables, 0.002 is the default gamma (1/n_variables)
# I only explicitly state it at this point so I can display it on graphs
gamma = 0.002
#compute kernels not preloaded into kpca
#laplacian
K_lap = laplacian_kernel(X_scaled, gamma=gamma)
#Import toy data and target
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,
y,
"Initial category 1 distribution of dataset %s" % dataset,
"Initial category 0 distribution of dataset %s" % dataset,
#output='show'
#output='plotly',
#ply_title="Initial distribution of dataset %s" % dataset,
output='save',
path='%sinitialdist.png' % filepath,
)
print('\nBoxplot of initial data for dataset %s saved.' % dataset)
## PREPROCESSING ##
#Scale initial data to centre data
X_scaled = scale(X)
X_scaled_df = pd.DataFrame.from_records(X_scaled)
#print(y)
#X2, y2 = target_split(inp_csv, 500)
print('\nShape of mvn dataframe: %s\n' % (X2.shape, ))
#print(X2)
print('\nShape of make_classification target array: %s\n' % (y2.shape, ))
#print(y2)
plt.figure(figsize=(50, 15))
plt.subplot(2, 1, 1)
img1 = sns.boxplot(data=inp_csv[inp_target == 1])
plt.title("Category A distribution of dataset 015", fontsize=20)
plt.subplot(2, 1, 2)
img2 = sns.boxplot(data=inp_csv[inp_target == 0])
plt.title("Category B distribution of dataset 015", fontsize=20)
#plt.savefig('../../data/simulated/mvnsim/mvnsim%sdist.png' % simname)
plt.show()
plt.close()
distribution_boxplot(
inp_csv,
inp_target,
"Category A distribution of dataset 015",
"Category B distribution of dataset 015",
output='show',
)