From e6e8840ac1e0654cd041cd186df0b0e6f971c2da Mon Sep 17 00:00:00 2001
From: Ricardo <acq11ra@sheffield.ac.uk>
Date: Fri, 14 Nov 2014 21:24:06 +0000
Subject: [PATCH] New file, sparse one vs all classification

---
 GPy/models/__init__.py                        |  1 +
 .../one_vs_all_sparse_classification.py       | 40 +++++++++++++++++++
 2 files changed, 41 insertions(+)
 create mode 100644 GPy/models/one_vs_all_sparse_classification.py

diff --git a/GPy/models/__init__.py b/GPy/models/__init__.py
index 6f0d5459..c6abb5de 100644
--- a/GPy/models/__init__.py
+++ b/GPy/models/__init__.py
@@ -20,3 +20,4 @@ from ss_mrd import SSMRD
 from gp_kronecker_gaussian_regression import GPKroneckerGaussianRegression
 from gp_var_gauss import GPVariationalGaussianApproximation
 from one_vs_all_classification import OneVsAllClassification
+from one_vs_all_sparse_classification import OneVsAllSparseClassification
diff --git a/GPy/models/one_vs_all_sparse_classification.py b/GPy/models/one_vs_all_sparse_classification.py
new file mode 100644
index 00000000..3db382ec
--- /dev/null
+++ b/GPy/models/one_vs_all_sparse_classification.py
@@ -0,0 +1,40 @@
+# Copyright (c) 2013, the GPy Authors (see AUTHORS.txt)
+# Licensed under the BSD 3-clause license (see LICENSE.txt)
+
+import numpy as np
+import GPy
+
+class OneVsAllSparseClassification(object):
+    """
+    Gaussian Process classification: One vs all
+
+    This is a thin wrapper around the models.GPClassification class, with a set of sensible defaults
+
+    :param X: input observations
+    :param Y: observed values, can be None if likelihood is not None
+    :param kernel: a GPy kernel, defaults to rbf
+
+    .. Note:: Multiple independent outputs are not allowed
+
+    """
+
+    def __init__(self, X, Y, kernel=None,Y_metadata=None,messages=True):
+        if kernel is None:
+            kernel = GPy.kern.RBF(X.shape[1])
+
+        likelihood = GPy.likelihoods.Bernoulli()
+
+        assert Y.shape[1] == 1, 'Y should be 1 column vector'
+
+        labels = np.unique(Y.flatten())
+
+        self.results = {}
+        for yj in labels:
+            print 'Class %s vs all' %yj
+            Ynew = Y.copy()
+            Ynew[Y.flatten()!=yj] = 0
+            Ynew[Y.flatten()==yj] = 1
+
+            m = GPy.models.SparseGPClassification(X,Ynew,kernel=kernel,Y_metadata=Y_metadata)
+            m.optimize(messages=messages)
+            self.results[yj] = m.predict(X)[0]