Skip to content

This software contains three closed tree mining algorithms: an incremental one IncTreeNat, a sliding-window based one, WinTreeNat, and finally one that mines closed trees adaptively from data streams, AdaTreeNat. Closed patterns are powerful representatives of frequent patterns, since they eliminate redundant information.

abifet/adatreenat

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 

Repository files navigation

AdaTreeNat

Mining Adaptively Frequent Closed Unlabeled Rooted Trees in Data Streams.

Albert Bifet and Ricard Gavaldà.

14th ACM SIGKDD Intl. Conference on Knowledge Discovery and Data Mining (KDD'08)

This software is an implementation of three closed unlabeled tree mining algorithms:

  • an incremental one, INCTREENAT
  • a sliding-window based one, WINTREENAT
  • one that mines closed trees adaptively from data streams, ADATREENAT

INSTALL:

To compile the program : type 'make' in the test directory To clean: type 'make clean' in the test directory

PLATFORM:

This program has been compiled using g++ 4.2 and tested on Linux.

LICENSE TERMS:

Under GNU Public License terms.

USAGE:

The usage of the program is: AdaTreeNat [Options], where Options are:

  • -s Set the absolute or relative minimum support (default 10%)
  • -ord Ordered Trees
  • -uno Unordered Trees (default)
  • -topdown Top-down subtrees
  • -induced Induced subtrees (default)
  • -in Set the name of the dataset (default 'input.trees')
  • -out Set the name of the output file (default 'output.trees')
  • -TreeNat TreeNat(default)
  • -inc IncTreeNat: Incremental Tree Mining
  • -win WinTreeNat:Sliding Window Tree Mining
  • -ada1 AdaTreeNat1:Adaptive Sliding Window Tree Mining
  • -ada2 AdaTreeNat2:Support ADWIN Tree Mining
  • -relaxed Uses relaxed support (default false)
  • -log Uses logarithmic relaxed support
  • -b Set size of the batches (default 10,000)
  • -w Set size of the sliding window (default 100,000)
  • -r Set the number of relaxed intervals (default 10,000)
  • -help Display this ussage help

The input should be a text file that uses:

  • TreeNat's string encoding form :

    • each line represents a transaction
    • the string is a depth-first traversal of the rooted tree where "[" represents a new node and "]" means a backtrack

    For example, here is a database with 3 rooted trees.

    [[[][][]][[][]]]

    [[[][][]][[[]]]]

    [[[[[]]]]]

  • Professor Zaki's string encoding form:

    • each line represents a transaction
    • the first two numbers are identical, they are tids
    • the 3rd number represents the length of the remaining string
    • the rest string is a depth-first traversal of the rooted tree, where "-1" means a backtrack.

    For example, here is a database with 3 rooted trees.

    0 0 15 1 1 1 -1 1 -1 1 -1 -1 1 1 -1 1 -1 -1

    1 1 15 1 1 1 -1 1 -1 1 -1 -1 1 1 1 -1 -1 -1

    2 2 9 1 1 1 1 1 -1 -1 -1 -1

About

This software contains three closed tree mining algorithms: an incremental one IncTreeNat, a sliding-window based one, WinTreeNat, and finally one that mines closed trees adaptively from data streams, AdaTreeNat. Closed patterns are powerful representatives of frequent patterns, since they eliminate redundant information.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages