GSP Algorithm is an Apriori based method with some enhancements.
After reading several description, I still could not figure out the enhancements brought by GSP in regards to the general Apriori algorithm. Is it the itemset order that is taken into account ?
Could you give me an example as I am a newbie in data mining.
Thank you in advance.
Apriori is to find frequent itemsets in transactions. A transaction is just an unordered set of items.
Apriori will output patterns that are set of items.
GSP is to find frequent sequential patterns in a sequences.
A sequence is an ordered list of transactions.
GSP will output patterns that are subsequences.
If you want to try Apriori and GSP, you can get the Java source code in the SPMF open source data mining library.
The difference between the two is that Apriori is for itemset mining and GSP is for sequence mining. It is based on Apriori but takes into account the order of the items and thus finds sequences. Hence abc is different from cba for example.
GSP is an Apriori-based method in the sequential pattern mining like AprioriAll. GSP add some properties which tends to solve AprioriAll limitations. They are "adding time constraint", "Sliding window time", and "taxonomies". You can find complete explanation here: http://www.philippe-fournier-viger.com/spmf/GSP96.pdf
