gsp_pattern_miner

GSP sequential pattern miner for sequence datasets. The library depends on the sequential_pattern_mining_protocols support library, implements the generic pattern_miner_protocol defined in the pattern_mining_protocols core library, and mines frequent sequential patterns using prefix-indexed candidate joins, pruning candidates whose immediate subpatterns are not all frequent, and counting singleton and candidate supports in batched horizontal dataset scans.

Requires a dataset implementing sequence_dataset_protocol with sequences represented as ordered lists of canonical sorted itemsets over a declared item domain.

API documentation

Open the ../../apis/library_index.html#gsp_pattern_miner link in a web browser.

Loading

To load this library, load the loader.lgt file:

| ?- logtalk_load(gsp_pattern_miner(loader)).

Testing

To test this library predicates, load the tester.lgt file:

| ?- logtalk_load(gsp_pattern_miner(tester)).

Features

  • Level-Wise Mining: Builds longer frequent sequences by extending the current frequent level.

  • Join-Based Candidates: Generates candidate sequences by joining compatible frequent patterns from the previous level.

  • Support-Pruned Candidates: Prunes candidates whose immediate subpatterns are not all frequent.

  • Batched Horizontal Support Counting: Counts singleton and candidate supports using horizontal scans over the sequence database at each level.

  • Canonical Sequences: Validates that itemsets are sorted, duplicate-free, non-empty, and restricted to declared items.

  • Flexible Support Thresholds: Supports relative minimum support and absolute minimum support count.

  • Model Export: Mined pattern collections can be exported as predicate clauses or written to a file.

Options

The mine/3 predicate accepts the following options:

  • minimum_support/1: Relative minimum support threshold in the interval ]0.0, 1.0]. The default is 0.5.

  • minimum_support_count/1: Absolute minimum support count. When both support options are provided, this option takes precedence.

  • maximum_pattern_length/1: Maximum total number of items in a mined sequential pattern. The default is 1000, effectively capped by the longest sequence in the dataset.

  • minimum_pattern_length/1: Minimum total number of items retained in the mined result. The default is 1.

Pattern miner representation

The mined pattern miner result is represented by a compound term with the functor chosen by the implementation and arity 3. For example:

gsp_pattern_miner(ItemDomain, Patterns, Options)

Where:

  • ItemDomain: Canonical sorted list of declared dataset items.

  • Patterns: List of sequence_pattern(Pattern, SupportCount) terms ordered first by total item count and then lexicographically.

  • Options: Effective mining options used to mine the frequent sequential patterns.

References

  1. Srikant, R. and Agrawal, R. (1996) - “Mining sequential patterns: Generalizations and performance improvements”.