BlackCurveNumpyLayer1

BlackCurveNumpyLayer1

@@ -0,0 +1,161 @@
+'''
+Created on Sep 1, 2018
+
+@author: Ashesh
+'''
+import numpy as np
+import pandas as pd
+from csv import reader
+from sklearn.model_selection import KFold
+
+# X  = (hours sleeping, hours studying), y = score on test
+'''
+X = np.array(([2, 9], [1, 5], [3, 6]), dtype=float)
+y = np.array(([92], [86], [89]), dtype=float)
+x = np.array(([4,8]), dtype=float)
+ytheta = np.array(([91]), dtype=float)
+'''
+'''
+X=pd.read_csv("train_theta.csv")
+X = X.values
+y=pd.read_csv("train_int.csv")
+y = y.values
+x=pd.read_csv("test_theta.csv")
+x = x.values
+ytheta=pd.read_csv("test_int.csv")
+ytheta = ytheta.values
+'''
+dataset = pd.read_csv("BlackCurveDataset.csv")
+dataset = dataset.values
+kf = KFold(n_splits=5, shuffle=True, random_state=None)
+i = 0
+for train_index, test_index in kf.split(dataset):
+    X_train, X_test = dataset[train_index], dataset[test_index]
+    print("X_train:", X_train, "X_test:", X_test)
+    if(i == 4):         #To set different partition of train and test dataSet
+        break
+    i +=1
+X = X_train[:,[0]]
+y = X_train[:,[1]]
+x = X_test[:,[0]]
+ytheta = X_test[:,[1]]    
+#print(len(X))
+#print(len(y))
+#print(len(x))
+#print(np.amax(y, axis=0))
+# scale units
+'''
+X = X/np.amax(X, axis=0) # maximum of X array
+x=x/np.amax(x,axis=0)
+y = y/22000 # max y value in the graph
+'''
+X = X/100 # maximum of X array
+x=x/100
+ymax = np.amax(y)
+y = y/np.amax(y, axis=0) # max y value in the graph
+ytheta=ytheta/np.amax(ytheta, axis=0)
+print(X.shape)
+dataSetFile2 = open("test_set.csv", "r+")
+dataSetFile3 = open("test_set_normalised.csv", "r+")
+for i in np.nditer(x):
+    #print(i)
+    dataSetFile2.write(str(i)+"\n")
+
+for i in np.nditer(ytheta):
+    #print(i)
+    dataSetFile3.write(str(i)+"\n")
+
+# Load a CSV file
+def load_csv(filename):
+    dataset = list()
+    with open(filename, 'r') as file:
+        csv_reader = reader(file)
+        for row in csv_reader:
+            if not row:
+                continue
+            dataset.append(row)
+    return dataset
+
+class Neural_Network(object):
+    def __init__(self):
+        #parameters
+        self.inputSize = 1
+        self.outputSize = 1
+        self.hiddenSize = 2000
+
+        #weights
+        self.W1 = np.random.randn(self.inputSize, self.hiddenSize)*0.01 # (3x2) weight matrix from input to hidden layer
+        self.W2 = np.random.randn(self.hiddenSize, self.outputSize)*0.01 # (3x1) weight matrix from hidden to output layer
+
+        #biases
+        #self.b1 = np.zeros((1, self.hiddenSize))
+        #self.b2 = np.zeros((self.outputSize, 1))
+
+    def forward(self, X):
+        #forward propagation through our network             
+        self.z = np.dot(X, self.W1)  # dot product of X (input) and first set of 3x2 weights
+        self.z2 = self.sigmoid(self.z) # activation function
+        self.z3 = np.dot(self.z2, self.W2) # dot product of hidden layer (z2) and second set of 3x1 weights
+        o = self.sigmoid(self.z3) # final activation function
+        return o 
+
+    def sigmoid(self, s):
+        # activation function 
+        return 1/(1+np.exp(-s))
+
+    def sigmoidPrime(self, s):
+        #derivative of sigmoid
+        return s * (1 - s)
+
+    def backward(self, X, y, o, l_rate):
+        # backward propgate through the network
+        self.o_error = y - o # error in output
+        self.o_delta = self.o_error*self.sigmoidPrime(o) # applying derivative of sigmoid to error
+        #self.b2_delta = (1/X.size)*np.sum(self.o_delta, axis=0, keepdims=True)
+
+        self.z2_error = self.o_delta.dot(self.W2.T) # z2 error: how much our hidden layer weights contributed to output error
+        self.z2_delta = self.z2_error*self.sigmoidPrime(self.z2) # applying derivative of sigmoid to z2 error
+        #self.b1_delta = (1/X.size)*np.sum(self.z2_delta, axis=0, keepdims=True)
+        #print(self.b1_delta.shape)
+
+        self.W1 += (l_rate/X.size)*(X.T.dot(self.z2_delta)) # adjusting first set (input --> hidden) weights
+        #self.b1 += (l_rate*self.b1_delta)
+
+        self.W2 += (l_rate/X.size)*(self.z2.T.dot(self.o_delta))  # adjusting second set (hidden --> output) weights
+        #self.b2 += l_rate*self.b2_delta
+
+    def train (self, X, y, l_rate):
+        o = self.forward(X)
+        self.backward(X, y, o, l_rate)
+
+    def saveWeights(self):
+        np.savetxt("w1.txt",self.W1,fmt="%s")
+        np.savetxt("w2.txt",self.W2,fmt="%s")
+
+    def predict(self):
+        print("predicted data:")
+        #print("input:\n"+str(x))
+        print("output:\n"+str(self.forward(x)*ymax))
+        # load and prepare data
+
+        dataSetFile1 = open("predicted.csv", "a")
+
+        for i in np.nditer(self.forward(x)):
+            print(i)
+            dataSetFile1.write(str(i*ymax)+"\n")
+
+
+NN = Neural_Network()
+for i in range(1000): # trains the NN 1,000 times
+    #print("Input: \n" + str(X)) 
+    #print("Actual Output: \n" + str(y)) 
+    #print("Predicted Output: \n" + str(NN.forward(X)))
+    print ("Loss: " + str(np.mean(np.square(y - NN.forward(X))))) # mean sum squared loss
+    print("\n")
+    l_rate = 0.01
+    NN.train(X, y, l_rate)
+
+NN.saveWeights()
+NN.predict()
+
+