COMP3270 Artificial Intelligence

Department of Computer Science

COMP3270 Artificial Intelligence

Assignment 3

1. Write a python program using sci-kit learn (aka sklearn) package to build a

decision tree for the mushroom dataset from the UCI Machine Learning Repository.

Split the dataset into 75% training and 25% test set. Write down the accuracy for

the test set.

2. Use Naive Bayes and Support Vector Machine to perform classification on a face

database, the Labeled Faces in the Wild (LFW ). The database can be download

from:

http://vis-www.cs.umass.edu/lfw/lfw-funneled.tgz (233MB)

Sample code is available from

https://scikit-learn.org/stable/auto examples/applications/

plot face recognition.html

Study the sample code. Try to extract eigenfaces (PCA on the images), perform

classification and report results using

(a) Naive Bayes Classifier

(b) Support Vector Machine (SVM ) using radial basis function (rbf ) kernel

Remove unnecessary code from the sample.

3. Use the same dataset (LFW ) as in Q2.

Apply a pretrained convolutional neural network, (using VGG16 ), and replace

the fully-connected layers with your own. Freeze the weights of the convolutional

layers and only train the new FC layer.

Sample code for using pre-trained VGG16 for another classification task is available

from: