{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,2]],"date-time":"2026-02-02T22:31:06Z","timestamp":1770071466759,"version":"3.49.0"},"reference-count":0,"publisher":"Slovenian Association Informatika","issue":"5","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IJCAI"],"abstract":"<jats:p>The spread of misinformation on online platforms has made fake news detection systems more essential and demanding. TriFactNet is introduced in this study as an innovative multi-factor deep learning approach combining semantic textual features, credibility of sources, and synthetic stance vectors for improved fake news accuracy and reliability. The model is trained and tested on an equal subset of the ISOT Fake News Detection Dataset consisting of 1,000 real and fake news articles labeled accordingly. For enhancing input representation, credibility of sources is synthesized using ground-truth ratings and 32-dimensional random stance vectors are added to mimic alignment of context with surrounding claims. The textual information is represented using lightweight transformer model\u2014prajjwal1\/bert-tiny\u2014while the auxiliary features are processed using parallel dense layers. These representations are combined and fed into fully connected layers for binary classification. The AdamW optimizer is used in training and ten epochs are used to test using accuracy as well as precision, recall, F1-score, and confusion matrix. Experimentation shows high performance in classification with overall accuracy being 97.5%, class-wise balanced metrics, and harmonized training-validation curves. Modular nature of the architecture and processing of multiple signals of information highlight its applicability to real-world disinformation detection. Future research will investigate the application of semantically derived stance vectors and large datasets to enhance scalability and generalizability.<\/jats:p>","DOI":"10.31449\/inf.v50i5.9272","type":"journal-article","created":{"date-parts":[[2026,2,2]],"date-time":"2026-02-02T10:25:45Z","timestamp":1770027945000},"source":"Crossref","is-referenced-by-count":0,"title":["TriFactNet: A Multi-Modal Neural Architecture for Fake News Detection Using Text, Source Credibility, and Stance"],"prefix":"10.31449","volume":"50","author":[{"given":"Jeena","family":"Joseph","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"16141","published-online":{"date-parts":[[2026,2,2]]},"container-title":["Informatica"],"original-title":[],"link":[{"URL":"https:\/\/www.informatica.si\/index.php\/informatica\/article\/download\/9272\/6432","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.informatica.si\/index.php\/informatica\/article\/download\/9272\/6432","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,2,2]],"date-time":"2026-02-02T10:25:45Z","timestamp":1770027945000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.informatica.si\/index.php\/informatica\/article\/view\/9272"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,2,2]]},"references-count":0,"journal-issue":{"issue":"5","published-online":{"date-parts":[[2026,2,2]]}},"URL":"https:\/\/doi.org\/10.31449\/inf.v50i5.9272","relation":{},"ISSN":["1854-3871","0350-5596"],"issn-type":[{"value":"1854-3871","type":"electronic"},{"value":"0350-5596","type":"print"}],"subject":[],"published":{"date-parts":[[2026,2,2]]}}}