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

Data science relies on pipelines that are organized in the form ofinterdependent computational steps. Each step consists of various candidatealgorithms that maybe used for performing a particular function. Each algorithmconsists of several hyperparameters. Algorithms and hyperparameters must beoptimized as a whole to produce the best performance. Typical machine learningpipelines consist of complex algorithms in each of the steps. Not only is theselection process combinatorial, but it is also important to interpret andunderstand the pipelines. We propose a method to quantify the importance ofdifferent components in the pipeline, by computing an error contributionrelative to an agnostic choice of computational steps, algorithms andhyperparameters. We also propose a methodology to quantify the propagation oferror from individual components of the pipeline with the help of a naive setof benchmark algorithms not involved in the pipeline. We demonstrate ourmethodology on image classification pipelines. The agnostic and naivemethodologies quantify the error contribution and propagation respectively fromthe computational steps, algorithms and hyperparameters in the imageclassification pipeline. We show that algorithm selection and hyperparameteroptimization methods like grid search, random search and Bayesian optimizationcan be used to quantify the error contribution and propagation, and that randomsearch is able to quantify them more accurately than Bayesian optimization.This methodology can be used by domain experts to understand machine learningand data analysis pipelines in terms of their individual components, which canhelp in prioritizing different components of the pipeline.

Quantifying contribution and propagation of error from computational steps, algorithms and hyperparameter choices in image classification pipelines