A novel hybrid model for identifying the most informative instances for improving text data classification

Amira Abdelwahab, Mohamed Salama · 2026

Read A novel hybrid model for identifying the most informative instances for improving text data classification by Amira Abdelwahab, Mohamed Salama (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0355603.txt.

Description

Amira Abdelwahab and Mohamed Salama propose a hybrid active learning system that picks which unlabelled texts are worth paying humans to annotate. Labelling is the bottleneck of text classification, especially in mixed, messy collections. Their 2026 framework pools four classic classifiers into an ensemble and loops: label a little, retrain, then ask for the most informative next examples.

Texts were converted to numbers four ways, from word counts to modern contextual embeddings, so representation choices could be compared fairly. Benchmarks came from healthcare, finance, spam, and shopping domains. Abdelwahab and Salama argue the ensemble-plus-selection loop reaches useful accuracy with far fewer labels than labelling everything, a practical gain for small teams.

Catalog record

CREATORAmira Abdelwahab, Mohamed Salama
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE54023
LICENSECC_BY
RIGHTSCC BY
SOURCEPLOS

Source: https://doi.org/10.1371/journal.pone.0355603

Browse the full library · About the Tower of Babel