English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Recent advances in time series classification have largely focused on methodsthat either employ deep learning or utilize other machine learning models forfeature extraction. Though successful, their power often comes at therequirement of computational complexity. In this paper, we introduce GeoStatrepresentations for time series. GeoStat representations are based off of ageneralization of recent methods for trajectory classification, and summarizethe information of a time series in terms of comprehensive statistics of(possibly windowed) distributions of easy to compute differential geometricquantities, requiring no dynamic time warping. The features used are intuitiveand require minimal parameter tuning. We perform an exhaustive evaluation ofGeoStat on a number of real datasets, showing that simple KNN and SVMclassifiers trained on these representations exhibit surprising performancerelative to modern single model methods requiring significant computationalpower, achieving state of the art results in many cases. In particular, we showthat this methodology achieves good performance on a challenging datasetinvolving the classification of fishing vessels, where our methods achieve goodperformance relative to the state of the art despite only having access toapproximately two percent of the dataset used in training and evaluating thisstate of the art.

GeoStat Representations of Time Series for Fast Classification