Top 10 Data Mining Algorithms

Find frequents items from a transaction dataset and derive association rules.
Figure 3 shows the Apriori Algorithm
Example
Suffers from a number of inefficiencies.
Improvements have been made. To learn more, take the data mining course (CS 530) next fall.

Published by Sergey Bring and Larry Page, the founders of Google, in 1998.
PageRank is a search ranking algorithm using hyperlinks on the Web. It is the basis for the Google search engine.
It is a static method that can be computed offline.
A hyperlink from page x to page y casts a vote. The value of the vote is based on the "importance" of page x.
Terminology: in-links of page i, out-links of page i.
Equation 12 shows the general idea.
Equation 17 shows the original idea in its full form. d is the damping factor and was set to 0.85
Figure 4 shows the power iteration method used to compute PageRank values.

Introduced by Freund and Shapire in 1995. They received the 2003 Gödel prize.
Ensemble learning method.
Solid theoretical foundation.
Accurate.
Simple.
Figure 5 shows the AdaBoost algorithm for a 2 class concept.
AdaBoost example
AdaBoost has been extended to multi-class learning problems.
A weak learning algorithm that performs just slightly better than random guessing can be boosted into an arbitrarily strong learning algorithm.