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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Adverse events are a serious issue in drug development and many predictionmethods using machine learning have been developed. The random splitcross-validation is the de facto standard for model building and evaluation inmachine learning, but care should be taken in adverse event prediction becausethis approach does not match to the real-world situation. The time split, whichuses the time axis, is considered suitable for real-world prediction. However,the differences in model performance obtained using the time and random splitsare not clear due to the lack of the comparable studies. To understand thedifferences, we compared the model performance between the time and randomsplits using nine types of compound information as input, eight adverse eventsas targets, and six machine learning algorithms. The random split showed higherarea under the curve values than did the time split for six of eight targets.The chemical spaces of the training and test datasets of the time split weresimilar, suggesting that the concept of applicability domain is insufficient toexplain the differences derived from the splitting. The area under the curvedifferences were smaller for the protein interaction than for the otherdatasets. Subsequent detailed analyses suggested the danger of confounding inthe use of knowledge-based information in the time split. These findingsindicate the importance of understanding the differences between the time andrandom splits in adverse event prediction and strongly suggest that appropriateuse of the splitting strategies and interpretation of results are necessary forthe real-world prediction of adverse events. We provide analysis code anddatasets used in the present study(https://github.com/mizuno-group/AE_prediction).

Investigation of a Data Split Strategy Involving the Time Axis in Adverse Event Prediction Using Machine Learning