Added SVM

code/artificial_intelligence/src/support_vector_machine/SVM.py

@@ -0,0 +1,51 @@
+
+
+import numpy as np 
+
+class svm:
+	def __init__(self):
+		self.b=None
+		self.t=None
+		self.count=None
+		self._lambda=None
+	def train(self,xtrain,ytrain,count,_lambda):
+		self.count=count
+		self.b=[]
+		for i in range(0,len(xtrain[0])):
+			self.b.append(0.0)
+		self.t=1
+		self._lambda=_lambda
+		for i in range(self.count):
+			for j in range(len(xtrain)):
+				output=ytrain[j]*xtrain[j][0]*self.b[0]
+				for k in range(1,len(xtrain[0])):
+					output=output+xtrain[j][k]*self.b[k]
+
+				if(output>1):
+					for k in range(0,len(xtrain[0])):
+						self.b[k]=self.b[k]*(1-(1/self.t))#/(self._lambda*self.t)
+				else:
+					for k in range(0,len(xtrain[0])):
+						self.b[k]=self.b[k]*(1-(1/self.t))+(ytrain[j]*xtrain[j][k])/(self._lambda*self.t)
+				self.t=self.t+1
+
+	def predict(self,xtest):
+		p_array=[]
+		for i in range(len(xtest)):
+			output=0
+			for j in range(len(xtest[0])):
+				output=output+self.b[j]*xtest[i][j]
+			if(output<0):
+				p_array.append(-1)
+			else:
+				p_array.append(1)
+
+		return p_array
+
+xtrain=[[2.327868056,2.458016525],[3.032830419,3.170770366],[4.485465382,3.696728111],[3.684815246,3.846846973],[2.283558563,1.853215997],[7.807521179,3.290132136],[6.132998136,2.140563087],[7.514829366,2.107056961],[5.502385039,1.404002608],[7.432932365,4.236232628]]
+ytrain=[-1,-1,-1,-1,-1,1,1,1,1,1]
+
+clf=svm()
+clf.train(xtrain,ytrain,16,0.5)
+print(clf.predict(xtrain))
+