Scribd
Upload
42 views

The University of The South Pacific: Lab 2 (Worth 4%)

1) This lab assignment covers evaluating machine learning algorithms and applying them to solve real-world problems. 2) Students will implement linear regression to analyze randomly generated 2D data and determine the best fit line. They will calculate the sum of squared errors and draw the best fit line with data points. 3) Students will then generate two clusters of random 2D data and use linear regression for classification, labeling the data classes with +1 and -1. They must determine the separation plane and test how well linear regression performs classification on the data.

Uploaded by

Donald Bennett
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
42 views

The University of The South Pacific: Lab 2 (Worth 4%)

1) This lab assignment covers evaluating machine learning algorithms and applying them to solve real-world problems. 2) Students will implement linear regression to analyze randomly generated 2D data and determine the best fit line. They will calculate the sum of squared errors and draw the best fit line with data points. 3) Students will then generate two clusters of random 2D data and use linear regression for classification, labeling the data classes with +1 and -1. They must determine the separation plane and test how well linear regression performs classification on the data.

Uploaded by

Donald Bennett
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

The University of the South Pacific

School of Computing, Information & Mathematical Sciences

CS412 Artificial Intelligence

Lab 2 (worth 4%)


This assessment covers the following course learning outcomes partially:
CLO 1 – Evaluate machine learning algorithms and their limitations
CLO 3 – Apply machine learning algorithms to solve real life complex computing
problems.

In this lab you will understand how linear regression works that you have learnt in your
lecture week 3. Do the following questions:

1 - Read and understand linear regression.


2 - Randomly generate data in 2D plane and determine the best fit line using linear
regression formula [W = pseudoinverse(X)*Y].
3 - Draw the best fit line together with data points. What is your sum of square Error
(Ein)?
4 - Test your program.
5 - Now generate random data as two separate clusters in 2D plane.
6 - Scatter plot your data. See the sample graphs below

Regression Classification

7 - Using linear regression for classification classify your data into two classes using the
same formula given above. [Note: this time you have to label your data with +1 and -1 to
indicate two different classes]. Does it give you an expected solution? Try different class
sizes.
8 - Draw the separation plane if possible. [You need to determine the linear equation
W'X=0 or Xw = 0]
9 - Test your program.
10 – Show how linear regression for classification can improve pocket algorithm with
PLA.
Unsatisfactory Satisfactory Good Marks % Marks
BOK
(0%-49%) (50% - 75%) (76% - 100%) Allocated Attained
Unsatisfactory Satisfactory Good
II. Irrelevant prototyping I. Analysis of I. All satisfactory and
VIII. Poor architecture and requirements such as …
design makes future risk analysis & II. Choice of architecture
updates hard to mitigation, and design is based on
incorporate. prototyping, robustness,
identification of scalability, evolution
Abstraction & true/false and long term
1
Design requirements etc. management.
III. all relevant III. Innovative or creative
constraints and design of algorithms
requirements are (efficiency, structure
mentioned and data structures)
VI. Appropriate choice
of the architecture
and design
Unsatisfactory Satisfactory Good
I. Plagiarism I. Able to write a I. All satisfactory and…
II. Poor indentation, hard to complex code for an demonstrate very good
read and follow the code open ended problem programming skills.
III. Lots of bugs and/or II. Use of highest
Programming errors standard of 1
IV. Program produces programming
unexpected output practices such as
VII. Hard coding of data in modularity with
the program. appropriate
VIII. Poor structure architecture, low

- Anurag Sharma
coupling, high
cohesion, standard
docs (eg. Javadoc)
etc.
III. Appropriate testing is
done.

Unsatisfactory Satisfactory Good


VII. Very limited or VII. Appropriate data has I. All satisfactory and
undesirable data has been been collected III. Extensive analysis is
Data & collected IX. logical inference is done on the
Information IX. Poor or incorrect derived from the collected/given data. 2
Management inference is derived from collected/given data IV. Exceptional Inference
the collected/given data is derived from the
data.

Sub Total &


4
comments

- Anurag Sharma

You might also like