CPT+: Decreasing the time/space complexity of the compact prediction tree

Ted Gueniche, Philippe Fournier-Viger*, Rajeev Raman, S. Tseng

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

30 Scopus citations

Abstract

Predicting next items of sequences of symbols has many applications in a wide range of domains. Several sequence prediction models have been proposed such as DG, All-k-order Markov and PPM. Recently, a model named Compact Prediction Tree (CPT) has been proposed. It relies on a tree structure and a more complex prediction algorithm to offer considerably more accurate predictions than many state-of-the-art prediction models. However, an important limitation of CPT is its high time and space complexity. In this article, we address this issue by proposing three novel strategies to reduce CPT’s size and prediction time, and increase its accuracy. Experimental results on seven real life datasets show that the resulting model (CPT+) is up to 98 times more compact and 4.5 times faster than CPT, and has the best overall accuracy when compared to six state-of-the-art models from the literature: All-K-order Markov, CPT, DG, Lz78, PPM and TDAG.

Original languageEnglish
Title of host publicationAdvances in Knowledge Discovery and Data Mining - 19th Pacific-Asia Conference, PAKDD 2015, Proceedings
EditorsTru Cao, Ee-Peng Lim, Tu-Bao Ho, Zhi-Hua Zhou, Hiroshi Motoda, David Cheung
PublisherSpringer Verlag
Pages625-636
Number of pages12
ISBN (Print)9783319180311
DOIs
StatePublished - 1 Jan 2015
Event19th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2015 - Ho Chi Minh City, Viet Nam
Duration: 19 May 201522 May 2015

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9078
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference19th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2015
CountryViet Nam
CityHo Chi Minh City
Period19/05/1522/05/15

Keywords

  • Accuracy
  • Compression
  • Next item prediction
  • Sequence prediction

Fingerprint Dive into the research topics of 'CPT+: Decreasing the time/space complexity of the compact prediction tree'. Together they form a unique fingerprint.

Cite this