Date of Award

Spring 2026

Degree Type

Honors College Thesis

Academic Program

Computer Science BS

Department

Computing

First Advisor

Dr. Nick Rahimi

Advisor Department

Computing

Abstract

This thesis proposes a multi-dimensional framework for news classification that evaluates articles across three independent dimensions: headline accuracy, language neutrality, and content reliability. These dimensions produce both a continuous reliability score and a five-tier interpretive scale, while additionally classifying articles by genre and topic. To operationalize this framework, a structured annotation protocol was developed and applied to a dataset of 373 news articles drawn from 79 outlets spanning a wide range of contemporary media ecosystem. A binary Logistic Regression classifier trained on the ISOT Fake News Dataset was then evaluated against this dataset to examine how a model trained on binary, source-narrow data performs on content of genuine complexity. The results demonstrate a substantial performance gap, with accuracy dropping from 95.62 % on ISOT to 69.71% on the multi-dimensional dataset, with approximately 70% of predictions falling below the confidence threshold and errors concentrating disproportionately in the most reliable and journalistically legitimate content. The analysis reveals that the model has learned a narrow stylistic template defined almost entirely by a single wire service source, rather than a generalizable conception of reliability. Furthermore, the genre dimension of the framework demonstrates that not all unreliable content is equivalent. Satire and misinformation are categorically distinct phenomena that binary frameworks conflate, with significant implications for automated content moderation. This work contributes both a conceptual framework and empirical evidence for the argument that reliable news classification requires instruments capable of representing reliability as the multidimensional, continuous, and context-dependent property it is.

Download

Share

COinS