A

According to Moroney, machine learning and traditional programming consist of three similar components: data, answers, and rules. The difference between these two types of programming is the order in which the components are used. In machine learning, data and answers are used as inputs to generate rules. In traditional programming, data and rules are used as inputs to generate answers.

B

import numpy as np 
import tensorflow
from tensorflow import keras

model = keras.Sequential([keras.layers.Dense(units = 1, input_shape = [1])])
model.compile(optimizer = 'sgd', loss = 'mean_squared_error')

xs = np.array([-1.0,0.0,1.0,2.0,3.0,4.0], dtype = float)
ys = np.array([-3.0,-1.0,1.0,3.0,5.0,7.0], dtype = float)

model.fit(xs,ys,epochs = 500)

print(model.predict([7.0]))

First result: 12.988924

Second result: 12.987505

Explaination: Every time the model is fit, the result will be slightly different because it starts with a randomized weight and the optimizer and loss function work methodically from that original weight. The final loss/accuracy of the model will be very similar but never exactly the same. As you can see above, the results for predicting 7.0 are the same for the first 4 digits and then vary due to slight variation in learning.

C

Best deal:

Worst deal:

Justification: