Description
Pr.2.1
In this assignment you will use the datasets from Project 1. In language of your preference (Python, Java,
Matlab, C++), implement a perceptron-based classifier that will iterate until the total error is:
Epsilon <10-5
, for dataset A,
Epsilon <10-1
, for dataset B,
Epsilon <5*10-1
, for dataset C.
To do this, you need to introduce a stopping criterion. You should also introduce a limit on maximum
number of iterations (let that be ni=5,000). Normalize the datasets first. Initialize your neuron using
random values between (-0.5, 0.5).
Please use unipolar version of:
a) Hard activation function
b) Soft activation function
For the scenario a) do the following for each of the datasets.
1. Choose 75% of the data for training, and the rest for testing. Train and test your neuron. Plot the
data and decision line for training and testing data (separately). Calculate errors for training and
testing dataset.
2. Choose 25% of the data for training, and the rest for testing. Train and test your neuron. Plot the
data and decision line for training and testing data (separately). Calculate errors for training and
testing dataset.
3. Compare 1. and 2. Are errors different and if so, why? What is the effect of different data set and
effect of different training/testing distributions? When would you use option 1 and when option 2
above? Comment and discuss.
Repeat steps 1. through 3. for scenario b).
CMSC 409: Artificial Intelligence
Fall 2019, Instructor: Dr. Milos Manic, http://www.people.vcu.edu/~mmanic
Project 2
Page 2 of 2 CMSC 409 Artificial Intelligence, Project 2
Important: The data set lists all male and then all female data points. Think about which data points you
should use for training and which for testing (i.e. algorithm will fail if trained on one type of patterns and
tested on another).
Pr. 2.2 Soft vs. hard activation function
Compare and discuss results when hard activation was used vs. when soft activation was used. Comment
for each training/testing distribution 1, 2, and 3.
This assignment may be updated!
——————————————————————
Note:
1. The code must be user friendly. The TA must be able to test it simply by executing the code.
2. Project deliverable should be a zip file containing:
a. Written report with answers to the questions above in word or pdf format.
b. Training/testing data sets as decided in Pr. 2.1 steps 1 & 2
c. Requested plots.
d. Source code.
3. Submit your zip file to Blackboard. Please name the zip file as GroupName_Project2.zip.
