Sequences of symbols can be used to represent data in many domains such as text documents, activity logs, customer transactions and website click-streams. Sequence prediction is a popular task, which consists of predicting the next symbol of a sequence, given a set of training sequences. Although numerous prediction models have been proposed, many have a low accuracy because they are lossy models (they discard information from training sequences to build the model), while lossless models are often more accurate but typically consume a large amount of memory. This paper addresses these issues by proposing a novel sequence prediction model named SuBSeq that is lossless and utilizes the succinct Wavelet Tree data structure and the Burrows-Wheeler Transform to compactly store and efficiently access training sequences for prediction. An experimental evaluation shows that SuBSeq has a very low memory consumption and excellent accuracy when compared to eight state-of-the-art predictors on seven real datasets.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.