CPCS433
Artificial Intelligence Topics
Assignment #1
Assign Date: Saturday, November 12, 2022
Due date: Sunday, November 15,2022
This assignment will involve data analysis and reporting on methods and results using Python
code. I recommend using Google Colaboratory or any other platform to implement your code.
You may also use IPython notebook or if you prefer Word or Latex to show snippets of your code
with a written description of what you did. Then, export the finished report as a PDF file and
upload it to the Blackboard. Please submit an only single electronic report or simply submit your
Google Collaboratory file. Please ensure that there is enough information in your report/code
report to understand what you did without examining the code itself.
Problem 1: Artificial Neural Network Regression Model
In this task, you will build an artificial neural network classifier. You will, in particular,
train a two-layer feed-forward neural network that predicts a student chance of
getting accepted in graduate school using three features (GRE_score, TOEFL _score,
and GPA). The data set has four examples for training and three examples for testing.
The following table shows the content of the dataset.
(Note that you are required to build the model from scratch without using existing
libraries).
Training Set:
GRE Score
TOEFL
GPA
300
308
329
332
99
103
110
118
6.8
8.36
9.15
9.36
GRE Score
TOEFL
GPA
296
293
325
95
97
112
7.54
7.8
8.96
Chance of
Acceptance
0.36
0.7
0.84
0.9
Testing Set:
Chance of
Acceptance
0.44
0.64
0.8
In this task you will do the following:
1. Build a two-layers neural network of one input layer (three nodes ??1 , ??2 and b or ??0 ),
one hidden layer (three nodes with sigmoid activation function) and one output layer
(one node with no activation function).
2. Initialize the parameters with some random values.
3. Perform a feed-forward stage to propagate the input forward through the network.
4. Preform the back-propagation algorithm using sum square error (SSE).
5. Print the content of the input/output of hidden layers and output layer at the feed
forward stages.
6. Print the content of the weight matrices after each epoch (4 epochs).
7. Test and report the result of classifying the testing examples.

Purchase answer to see full
attachment