You are here

Frequent Pattern Mining for Microarray Data

NeSC Research Seminar Series
Speaker: 
Andrei Lyashko

Frequent Pattern Mining (FPM) represents a very recent field of data mining and an evolutionary technique to ARM. Recalling an example from ARM, frequent pattern is simply a set {A, C} with no implication and is therefore only categorised by the support, which is basically the number of joint occurrences of all elements in the set. Although it may seem as a downgrade in information that it represents, frequent patterns represent a huge upgrade in efficiency of mining and compactness of its representation, while maintaining all the required information to derive association rules on request. Moreover, a compact version called frequent closed patterns exists that represents all relationships between items in a very condensed and lossless form which imply that, if successful, all relationships from the microarray data may be mined with the ability to later derive association rules from them.

Date and time: 
Wednesday, 5 August, 2009 - 10:00
Length: 
45 minutes
Location: 
Cramond