Skip to content

Releases: fidelity/seq2pat

seq2pat 1.4.0

13 Apr 15:19
e9e002f
Compare
Choose a tag to compare

Major:

  • Enable running Seq2Pat on batches of sequences in parallel
  • Add a usage example notebook to show how batch processing is called
  • Enrich documentations and comments relevant to batch processing and a few tests
  • Explicitly require that input integer sequences only contain positive integers

seq2pat 1.3.4

14 Jul 18:14
08f4091
Compare
Choose a tag to compare

Minor:

  • Fix a build error in release 1.3.3, which is reported by using the Visual Studio build tools

seq2pat 1.3.3

05 Jul 16:00
0a0fc60
Compare
Choose a tag to compare

Major:

  • Add a built-in maximum span constraint on items' indices to simplify and speed up usages
  • A fix to use original gap value instead of abs gap value in Check_gap() in build_mdd.cpp
  • Update unit tests where the built-in constraint may change the mining results

seq2pat 1.3.2

24 May 17:55
b36cfd5
Compare
Choose a tag to compare

Minor:

  • Refactor encodings generation component as Pat2Feat

seq2pat 1.3.1

28 Apr 21:23
8f545b7
Compare
Choose a tag to compare

Minor:

  • Set the default behavior of drop_pattern_frequency in encoding to be True

seq2pat 1.3.0

25 Apr 21:30
9f01dab
Compare
Choose a tag to compare

Major:

  • Add one-hot encodings, by mining patterns locally (csp_local.py) or globally
    (csp_global.py) in each sequence
  • Add Dichotomic Pattern Mining notebook, running on sample data
  • Clean up variables in backend