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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

In order to allow machine learning algorithms to extract knowledge from rawdata, these data must first be cleaned, transformed, and put intomachine-appropriate form. These often very time-consuming phase is referred toas preprocessing. An important step in the preprocessing phase is featureselection, which aims at better performance of prediction models by reducingthe amount of features of a data set. Within these datasets, instances ofdifferent events are often imbalanced, which means that certain normal eventsare over-represented while other rare events are very limited. Typically, theserare events are of special interest since they have more discriminative powerthan normal events. The aim of this work was to filter instances provided tofeature selection methods for these rare instances, and thus positivelyinfluence the feature selection process. In the course of this work, we wereable to show that this filtering has a positive effect on the performance ofclassification models and that outlier detection methods are suitable for thisfiltering. For some data sets, the resulting increase in performance was only afew percent, but for other datasets, we were able to achieve increases inperformance of up to 16 percent. This work should lead to the improvement ofthe predictive models and the better interpretability of feature selection inthe course of the preprocessing phase. In the spirit of open science and toincrease transparency within our research field, we have made all our sourcecode and the results of our experiments available in a publicly availablerepository.

Outlier Detection as Instance Selection Method for Feature Selection in Time Series Classification