Computer Science > Machine Learning
[Submitted on 2 Oct 2020]
Title:Modifying the Symbolic Aggregate Approximation Method to Capture Segment Trend Information
View PDFAbstract:The Symbolic Aggregate approXimation (SAX) is a very popular symbolic dimensionality reduction technique of time series data, as it has several advantages over other dimensionality reduction techniques. One of its major advantages is its efficiency, as it uses precomputed distances. The other main advantage is that in SAX the distance measure defined on the reduced space lower bounds the distance measure defined on the original space. This enables SAX to return exact results in query-by-content tasks. Yet SAX has an inherent drawback, which is its inability to capture segment trend information. Several researchers have attempted to enhance SAX by proposing modifications to include trend information. However, this comes at the expense of giving up on one or more of the advantages of SAX. In this paper we investigate three modifications of SAX to add trend capturing ability to it. These modifications retain the same features of SAX in terms of simplicity, efficiency, as well as the exact results it returns. They are simple procedures based on a different segmentation of the time series than that used in classic-SAX. We test the performance of these three modifications on 45 time series datasets of different sizes, dimensions, and nature, on a classification task and we compare it to that of classic-SAX. The results we obtained show that one of these modifications manages to outperform classic-SAX and that another one slightly gives better results than classic-SAX.
Submission history
From: Muhammad Marwan Muhammad Fuad [view email][v1] Fri, 2 Oct 2020 00:05:39 UTC (720 KB)
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
IArxiv Recommender
(What is IArxiv?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.