JNTU B.Tech IV DATA ANALYTICS Previous Question Paper (September 2020 - R16)

To identify outliers, calculate the Interquartile Range (IQR). First, sort the data: {10, 13, 15, 19, 22, 25, 36, 40, 88, 94}. With n=10, the median is (22+25)/2 = 23.5. The first quartile (Q1) is the median of the lower half (10, 13, 15, 19, 22), which is 15. The third quartile (Q3) is the median of the upper half (25, 36, 40, 88, 94), which is 40. The IQR is Q3 - Q1 = 40 - 15 = 25. The lower bound is Q1 - 1.5 * IQR = 15 - 37.5 = -22.5. The upper bound is Q3 + 1.5 * IQR = 40 + 37.5 = 77.5. Values 88K and 94K exceed the upper bound and are therefore outliers.

Outliers significantly inflate measures of spread like the range and standard deviation. While the interquartile range remains robust, the mean and standard deviation become skewed, failing to represent the central tendency of the majority of the data points, which are clustered below 40K.

Missing values can be treated using several techniques:

• Deletion: Removing rows or columns with missing data. For example, if a student record lacks a salary value, the entire row is dropped. This is effective only if the data is missing at random and the sample size is large.
• Mean/Median/Mode Imputation: Replacing missing values with the central tendency of the feature. For a dataset of salaries, replacing a missing value with the median (e.g., 22K) preserves the sample size but reduces variance.
• K-Nearest Neighbors (KNN) Imputation: Estimating missing values based on the similarity of other features. If a student has a missing salary, the system finds the 'k' most similar students based on GPA and skills and averages their salaries.
• Regression Imputation: Using a model to predict the missing value based on other correlated variables.