Description: The realization of GSP is the sequential mode analysis algorithm
GSP is an algorithm of sequential pattern mining. Its main description is as follows:
According to the length of the seed set Li, the candidate sequence mode Ci+1 with length I +1 is generated by connecting operation and shearing operation. Then scan the sequence database, calculate the support number of each candidate sequence pattern, generate the sequence mode of length I + 1, Li+ 1, and use Li+1 as the new seed set.
Repeat step 2 until no new sequence pattern or new candidate sequence pattern is produced.
L scan sequence database, and get the sequence pattern L1, length 1, as the initial seed set
L1Þ 2 Þ 3 Þ 4 & THORN.
There are two main steps to generating candidate sequence patterns
L stage connection: if you remove sequence model s1 sequence of the first item and remove the last item of s2 sequence generated by the same, you can connect s1 to s2, the s2 of the last item added to the s1.
L shear stage: if a subsequence of a candidate sequence pattern is not a sequence pattern, then this candidate sequence pattern cannot be a sequence pattern, which is removed from the candidate sequence pattern.
Candidate sequential patterns support calculation: for a given candidate set of sequential patterns, C scan sequence database, d for each of these sequences, find out all candidate set C be d contains sequential patterns, and increase the support count.
- [FP-GROWTH] - I upload the FP-GROWTH algorithm Visual
- [DBSCAN_JAVA] - DBSCAN algorithm JAVA, the D : \ text.tx
- [gsp] - Data Mining Association Rules classic al
- [GSP] - An important data mining algorithms, we
- [gsp] - Have a test based on a candidate sequent
- [45665985tem] - DS18B20 measurement temperature, and whe
- [www2008paper] - This is the annual www2008 all published
- [clospan] - Data Mining Sequential Pattern Mining Al
- [FPGROWTH] - Based on Improved FP-Tree of sequential
File list (Check if you may need any files):
Download