optimizing for case 1 and 2.

optimize.py

def parameter_optimize(x1, x2, y, w1=w1,w2=w2, b=b, learning_rate = learning_rate):
    # X contains the 100 student records.
    # Iterate through each record.
    for i in range(len(x1)):
        # Make prediction using the initial values of W[0], W[1], b.
        y_hat = find_perceptron_prediction(x1[i], x2[i], w1, w2, b) 
        # Case where the red points are wrongly classified.
        # This is the case where the actual output is 0 but the prediction is 1.
        if y[i] != y_hat and y[i] == 0:
            w1 = w1 - test_scores[i] * learning_rate
            w2 = w2 - grades[i] * learning_rate
            b  = b - learning_rate
        # Case where the green points are wrongly classified.
        # This is the case where the actual output is 0 but the prediction is 1.    
        if y[i] != y_hat and y[i] == 1:
            w1 = w1 + test_scores[i] * learning_rate
            w2 = w2 + grades[i] * learning_rate
            b = b + learning_rate
    return w1, w2, b