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/

Zendy Plus

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

Zendy Tools

Zendy Open

Machine learning (ML) over relational data is a booming area of the databaseindustry and academia. While several projects aim to build scalable and fast MLsystems, little work has addressed the pains of sourcing data and features forML tasks. Real-world relational databases typically have many tables (often,dozens) and data scientists often struggle to even obtain and join all possibletables that provide features for ML. In this context, Kumar et al. showedrecently that key-foreign key dependencies (KFKDs) between tables often lets usavoid such joins without significantly affecting prediction accuracy--an ideathey called avoiding joins safely. While initially controversial, this idea hassince been used by multiple companies to reduce the burden of data sourcing forML. But their work applied only to linear classifiers. In this work, we verifyif their results hold for three popular complex classifiers: decision trees,SVMs, and ANNs. We conduct an extensive experimental study using bothreal-world datasets and simulations to analyze the effects of avoiding KFKjoins on such models. Our results show that these high-capacity classifiers aresurprisingly and counter-intuitively more robust to avoiding KFK joins comparedto linear classifiers, refuting an intuition from the prior work's analysis. Weexplain this behavior intuitively and identify open questions at theintersection of data management and ML theoretical research. All of our codeand datasets are available for download fromhttp://cseweb.ucsd.edu/~arunkk/hamlet.

Are Key-Foreign Key Joins Safe to Avoid when Learning High-Capacity Classifiers?