Study plan
Introduction to Core Data Mining Concepts and Cautious Interpretation
Build a four-lesson introduction to the indexed core material in data mining, moving from data foundations to major mining tasks and cautious interpretation.
Lessons in this plan
-
1
Data Foundations: Objects, Attributes, and Proximity Measures
Learn how data objects are represented by attributes and how proximity is quantified using common distance measures such as Euclidean, Manhattan, Hamming, and supremum distances, providing the foundation for later mining tasks.
- 2 Classification Basics: Separating Classes and Generalization Understand the core goal of classification, including binary classification with separating hyperplanes, linear separability, and why choosing models for good generalization matters beyond zero training error.
- 3 Major Pattern Discovery Tasks: Association and Sequential Patterns Introduce two central pattern mining tasks by studying frequent itemsets, support, confidence, confidence-based pruning, and the role of constraints such as window size and maxgap in sequential pattern discovery.
- 4 Clustering and Careful Interpretation of Mining Results Examine clustering as a major unsupervised task, including similarity matrices, algorithms such as K-means and DBSCAN, computational considerations like spatial access methods, and cautions about finding apparent patterns in random data and making many statistical tests without inflating errors.