• University of Minnesota-Twin Cities

• CSCI 5523

CSCI 5523, Fall 2021 Midterm 1 Oct. 20th, 2021 2:30 PM - 3:45 PM Question 1. (7 points) 1. Discrete, Ordinal. 2. Continuous, Ratio. Question 2. (12 Points) a. True. Outliers will distort the mean and variance. b. False. The correlation is undefined since the variance is 0 for a constant vector. (Note 0/0 is undefined and does not equal 0!) c. False. Sampling would pick values at random and could miss entire periods of time, e.g., weeks or months, while aggregation can be done in a way that preserves the structure of the data, e.g., aggregating by weeks or months. Question 3. (8 points) a) The attributes have different scales and means. This would distort the distance measure. b) We can standardize each attribute using z-normalization so that their ranges are comparable. We can standardize each attribute by subtracting the mean and dividing by the standard deviation. Question 4. (15 points) a) Ripper b) Naïve Bayes c) Decision Trees