# -*- coding: utf-8 -*-
"""
'preprocessing_demo.py'

Data Standardization, or mean removal and variance scaling

David Pan, UAH
"""

import numpy as np
# Need to change the path name for the infile below
infile = "mvnrnd.csv"
dataset = np.loadtxt(infile, delimiter=',')
X = dataset[:, 0:2]
y = dataset[:,2]

from sklearn import preprocessing
scaler = preprocessing.StandardScaler().fit(X)

# Mean
scaler.mean_
X.mean(axis = 0)

# Standard deviation
scaler.scale_
X.std(axis = 0)

X_scaled = scaler.transform(X)
X_scaled.mean(axis = 0)
X_scaled.var(axis = 0)



from sklearn.neighbors import KNeighborsClassifier
neigh = KNeighborsClassifier(n_neighbors=5)
neigh.fit(X_scaled, y)

# Apply the same scaling on the test data Xin
Xin = X.mean(axis = 0)
Xin_scaled = scaler.transform([Xin])

neigh.predict(Xin_scaled)
neigh.predict_proba(Xin_scaled)

# Alternatively, use pipeline

from sklearn.pipeline import make_pipeline
from sklearn.neighbors import KNeighborsClassifier
from sklearn.preprocessing import StandardScaler


pipe = make_pipeline(StandardScaler(), KNeighborsClassifier(n_neighbors=5))

pipe.fit(X, y)
pipe.predict([Xin])
pipe.predict_proba([Xin])