#Kernel-Nearest-Neighbor-Algorithm-in-Python- This code extended to the well-known nearest-neighbor algorithm for classification so that kernels can be used

the distance is calculated by the following formula Distance=k(x,x)+k(y,y)-2k(x,y)

#Object obj=KernelKNNClassifier(Kernel) Kernel:name of kernal Parameters: kernal parameters in not parameters just put 0 see kernel type of more info

#Methos fit(y,X) y:vector of labels X: vector of features with columns corresponding to sample and rows corresponding to features
predicte(self,Z,NaborsNumber) Z:Feature matrix of vectors to be predicted. Rows correspond to sample index
NaborsNumber: number of nabors for Majority vote

#Example Knn=KernelKNNClassifier(['sigmoid'],[1,1]) X = np.array([[1, 1], [0, 1],[1, 0], [-1, -1], [1, -1],[-1, 0],[100,100]]); y=[1,1,1,0,0,0,2] Knn.fit(y,X) two Nabors Nabors=2 yhat=Knn.predicte(X, Nabors)

#Kernel references
1)Type:'rbf' Parameters=[gamma] Formula:K(x, y) = exp(-gamma ||x-y||^2)

2)Type:'sigmoid' Parameters=[ gamma , c0] Formula:K(x, y) = tanh(gamma <X, Y>+c0) gamma : slope c0 is known as intercep

3)Type:‘polynomial’ Parameters=[gamma,coef0,degree] Formula:K(X, Y) = (gamma <X, Y> + coef0)^degree coef0 : int, default 1 degree : int, default 3 3)Type:‘lin’ Formula:K(X, Y) = <X, Y>

4)Type:‘cosine’ L2-normalized dot product of vectors. K(X, Y) = <X, Y>/||X||||Y||