DSS4M Exam1 2019

Page 1

EXAM: Data Science and Security for Mobility Year: 2019/20, Semester: 1, Exam Paper: 1 (16/1/2020) Student ID: Name: Answer the questions in the space provided. If you run out of room, you can use the spare pages at the end, (but remember to clearly mark which question you are answering). 1. Data Mining is sometimes defined as the process of discovering patterns in data. What properties should those patterns have? 2. What is the difference between “supervised learning” and “unsupervised learning”? 3. List five important steps in a typical data science project. 4. If a temperature attribute takes values such as 27 degrees Celsius, is it a categorical, ordinal or numeric attribute? Would it make sense to use the ratio of two temperature values as a feature? Why or why not? 1 5. When building a regressor or classifier, how would one deal with an attribute “fashion” that takes values{versace, prada, armani, etc.}? Will there be an effect on the amount of memory required to store the training data? 6. What is a cyclic variable? Give an example of one. 7. What type of variable is shirt-size if it takes the values: {XS, S, M, L, XL, XXL}? 8. If an attribute is missing completely at random (MCAR), would removing all rows for which the attribute is missing affect the distribution of data in the training set? Explain. 9. If the training data contains very few missing values for a particular attribute would it make more sense to delete the rows for which the data is missing or the column? Why? Page 2 10. What if most of the values for a particular attribute were missing? Would it make more sense to delete the rows containing the missing values or the column? Why? 11. The frequency and mode of categorical attributes, and the mean and median of numeric attributes are examples of what…