JAWAHARLAL NEHRU TECHNOLOGICAL UNIVERSITY HYDERABAD MCA II DATA MINING Previous Question Paper (March - 2023 - R20)

Data smoothing is a technique used to remove noise from data to highlight important patterns. Common methods include:

• Binning: Data is sorted and partitioned into equal-width or equal-depth bins. Smoothing is then performed by replacing values with the bin mean, median, or boundaries.
• Regression: Data is fitted to a function (e.g., linear or multiple regression) to estimate the relationship between variables, effectively filtering out random noise.
• Clustering: Data points are grouped into clusters, and outliers are identified and removed or replaced based on their distance from cluster centroids.

A closed frequent itemset is an itemset where none of its immediate supersets have the same support count as the itemset itself. It provides a compressed representation of frequent itemsets without losing information.

A maximal frequent itemset is an itemset where none of its immediate supersets are frequent. This is a more aggressive compression technique; while it identifies the boundaries of frequent patterns, it does not retain the support counts of the subsets.