Individual Project 2 – BUSI 650
Fall 2021 – Dr. Said Baadel
Time limit: 24hrs
Student Name: Student ID:
Answer all Questions on the spaces provided. This assignment should be individual work. Once completed, save the file as a PDF and submit.
Use the Adult and Adolescent datasets provided for this exercise. See the grading ru
ic provided below for each question.
The datasets provided are for Adults and Adolescent participants who were surveyed as to whether they had symptoms of Autism[footnoteRef:1]. The survey is based on AQ-10 screening questionnaire that is used by the University of Cam
idge autism research center as a refe
al guide. Use the two datasets to compare the classifiers in ML classification algorithms and recommend which ones are best for predicting Autism Spectrum Disorder (ASD) traits. [1: Thabtah, F XXXXXXXXXXMachine learning in autistic spectrum disorder behaviour research: A review and ways forward. Informatics for Health and Social Care. DOI: XXXXXXXXXX/ XXXXXXXXXX1399132]
1. Draw a confusion matrix for the ASD test (2 Mark).
2. List the various Performance measures formulas to be used (i.e., Sensitivity, Specificity, Accuracy, F1 measure, and Precision) (2.5 Marks)
3. Draw 1 table highlighting the performance of the following classifiers (RIPPER, PART, Random Forest, Random Tree, Artificial Neural Network) for both the Adult and Adolescent datasets. In this table, show the performance measures listed in question 2 above. (7.5 Marks)
4. The datasets have the following instances of Family History of ASD.
Are there any differences in the performance of the classifiers with the Adult dataset compared to the Adolescent dataset? What can you attribute this difference to (if any)? What ways in ML are there to remedy such an issue? Hint: refer to Family History of ASD table above. (3 Marks)
5. Analyze the results in question 3 above. Explain in detail the performance of the above classifiers and highlight which one is better suited for these datasets based on what measure and why. (5 Marks)
4