

In [1]:

    
import numpy as np
import matplotlib as mpl
import matplotlib.pyplot as plt

from GPs.Kernel import Kernel, SqExp, RQ, ExpSine
from GPs.GP import GPCB, GPC

%matplotlib inline

Binary Classification (using GPCB)

Generate Training Data



In [2]:

    
train_size = 50
rng = np.random.RandomState(0)
x = rng.uniform(0, 5, 100)
y = np.array((x > 2.5)*2-1, dtype=int)
plt.scatter(x,y)









    Out[2]:





<matplotlib.collections.PathCollection at 0x7f5f1f9c1c50>

Train the Model



In [3]:

    
k1 = SqExp(1,1)
gpcb = GPCB(k1)
gpcb.train(x,y)

Predict



In [4]:

    
x_star = x

pi_hat_star_mean = gpcb.predict(x_star)
pi_star_mean = gpcb.predict(x_star,False)



In [5]:

    
plt.scatter(x_star,pi_hat_star_mean)
plt.scatter(x_star,pi_star_mean)









    Out[5]:





<matplotlib.collections.PathCollection at 0x7f5f224965c0>

Binary Classification (using GPC)

Generate Training Data

Redo training data such that the value of y is 1 if it belongs to class C and 0 otherwise, where y is a vector of length Cn for n training points. The first n points are for class 1, the next n points are for class 2, and so on.



In [6]:

    
x_t = x
y_t = np.array((x_t > 2.5), dtype=int)



In [7]:

    
x = np.append(x_t,x_t)



In [8]:

    
y_c1 = y_t #1 if it belongs to class 1, 0 otherwise
y_c2 = (y_t-1)*-1 #1 if it belongs to class 2, 0 otherwise

y = np.append(y_c1,y_c2)

Train the Model



In [9]:

    
k1 = SqExp(1,1)
gpc= GPC(k1)
gpc.train(x,y,2)



In [10]:

    
x_star = x_t
pi_star_mean_1 = gpc.predict(x_star,1)
pi_star_mean_2 = gpc.predict(x_star,2)



In [11]:

    
plt.scatter(x_star,pi_star_mean_1)
plt.scatter(x_star,pi_star_mean_2)









    Out[11]:





<matplotlib.collections.PathCollection at 0x7f5f1e412400>



In [12]:

    
lml = gpc.lml()
#print(lml)