Scribd
Upload
182 views

Assignment 2

This document provides instructions for an assignment on analyzing the Titanic dataset using decision trees. Students are asked to: 1) explore and preprocess the dataset to create new attributes; 2) load the preprocessed data into Weka and visualize distributions; 3) build a decision tree model on the training set and test it on the test set; 4) analyze and discuss the results compared to actual outcomes of the Titanic incident. The assignment must be submitted in a professional report with all code and outputs.

Uploaded by

Erick Menjivar
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
182 views

Assignment 2

This document provides instructions for an assignment on analyzing the Titanic dataset using decision trees. Students are asked to: 1) explore and preprocess the dataset to create new attributes; 2) load the preprocessed data into Weka and visualize distributions; 3) build a decision tree model on the training set and test it on the test set; 4) analyze and discuss the results compared to actual outcomes of the Titanic incident. The assignment must be submitted in a professional report with all code and outputs.

Uploaded by

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

CST8390 Assignment 2

Due: June 26, 2022 at 11:59 PM Sharp!!!


(Late submissions will not be accepted)
Goal: The goal of this lab is to explore and analyze Titanic dataset and perform classification using
Decision Trees.
References:
1. http://web.stanford.edu/class/archive/cs/cs109/cs109.1166/problem12.html
2. https://towardsdatascience.com/predicting-the-survival-of-titanic-passengers-
30870ccc7e8
3. https://www.kaggle.com/c/titanic
4. http://csis.pace.edu/~ctappert/srd2014/d3.pdf
5. https://titanicfacts.net/titanic-survivors/
Steps:
Data Understanding
Explore and analyse Titanic dataset given with this assignment (both train & test sets
provided). Read the pages given in references. You must include a brief description (what
is this dataset about, what is the purpose of analysing it etc. - 10 lines) about the dataset.
Also, you must provide a table with all attribute names and their description.
Data Preparation
1. Identify and record relevant attributes to perform a classification on this dataset. Remove
irrelevant attributes in the dataset.
2. Create a new attribute to represent age group
(If age is not given, keep it is as NK,
if age <2, then Baby;
else if age < 12, then Child;
else if age < 18, then Youth;
else if age ≤ 60, then Adult,
else if age >60, then Senior).
3. Create a new attribute “Relatives”. To create this column, the total number of relatives that
include siblings, spouse, parents and children should be calculated.
if the number of relatives is 0, record it as “None”,
else if the number is less than 3, record it as “Few”,
else if the number is less than 5, record it as “Some”
else if the number is 5 or greater, record it as “Many”.
4. Save the new file as Titanic_train_processed.csv. Provide a screenshot of this file (header
and a few rows should be visible.)
Load data into Weka. Double check the type of your attributes. If they are not as expected, apply
filters to convert them to the right types. (For example, class attribute should be nominal. Also, as
you will be using decision trees, it is meaningful to have nominal attributes)
Paste a screenshot of (a) distribution of the class attribute and (b) for the attribute Age Group by
selecting class attribute from the dropdown list for visualization. Include screen shots in the report.
Save the file as titanic_train_processed.arff. Include a screenshot of the file. (Open the file in
Notepad++. header and a few instances should be visible.)
Modeling & Evaluation
Prepare the test set in the same way that you prepared training set. Test file should have the same
format as the train set. In addition to all the other preprocessing steps, create a column for Survived,
with ‘?’ as the value. Save it as an arff file. Then use the same header from training file in the test
file too (only change relation name). Now, perform classification using Decision Trees with 10-
fold cross validation. Copy your confusion matrix and include it in your report. Visualize tree and
paste the tree. Explain the meaning of the obtained tree.
Now open another explorer and open your test file. This is just to ensure that the test file is in the
right format to be used for testing. If there is an issue in opening the file, you need to make changes
in the test file. Once test file is opened in the new explorer window, close the window. Now, in
the first explorer window, set the test set for testing. Click on Supplied Test set and set the test set.

Run Decision trees for the test set. As there is no actual Survived information, you will not get a
valid confusion matrix. Right click on the execution and visualize classifier errors. Save your file
from there as res.arff. Include a screenshot of this file. Your new file will have a new column
named “predicted Survived”. Fill in the following information:
a. Total instances in the test file:
b. Number of persons predicted to survive (1):
c. Number of persons predicted not to survive (0):
d. Percentage of predicted survival:
Discussion of Results
From reference 5, check the actual information of the incident. Give an explanation on how your
predicted results matches with the actual incident. List a few reasons why you think that your
answer is different from the actual results. Also, you need to compare results in detail based on
various features. This is a 5 marks question, so a detailed analysis and comparison of results
with tables and charts expected.

Note: Make sure that you have selected relevant attributes. Otherwise, the analysis
will be completely wrong and if you select totally irrelevant attributes that do not
have any effect on the survival, you will not get any marks.
Submission Details:
This is a partner assignment. Report should have a cover page with the names (Last name, first
name) and student numbers. You must paste all screenshots in the report. The report should have
table of contents, images, etc., sections titled Introduction, Data Understanding, Data Preparation,
Modeling and Evaluation, Discussion of Results, Conclusion, References etc.
Font: Times New Roman. Font size: 12, justified
Now, create a zipped folder named:
<LastNameFirstStudent>_<FirstNameFirstStudent>_<LastnameSecondStudent>_<FirstNameSe
condStudent>.zip with the
• Report in professional style,
• processed train and test arff files,
• model files and
• res.arff file.
Upload the zipped folder to Brightspace.

Marks:
This assignment will have a total of 30 marks. There will be negative marks if you miss
explanation for any of the steps. Every step/question should be answered with explanation.
Prepare your assignment in a professional report style.

You might also like