Comparison of basic recommendation methods for Czech news articles

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F61988987%3A17310%2F24%3AA25039G3" target="_blank" >RIV/61988987:17310/24:A25039G3 - isvavai.cz</a>

Výsledek na webu

<a href="https://ieeexplore.ieee.org/document/10900839" target="_blank" >https://ieeexplore.ieee.org/document/10900839</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/Informatics62280.2024.10900839" target="_blank" >10.1109/Informatics62280.2024.10900839</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Comparison of basic recommendation methods for Czech news articles

Popis výsledku v původním jazyce

The common contemporary recommendationmethods were assessed for recommending relevant news articlesin the Czech language. We used a TF-IDF vectorization withcosine similarity for content-based recommendations andsingular value decomposition for collaborative filtering. Thesimplest model was the popularity model based on the votecounts of other users. While the best days of expert systems seemto be gone in the era of more sophisticated neural networkmodels, we also wanted to investigate whether expert systemscan still show some benefits in providing relatively simple andexplainable methods for an ensemble of recommendations. Wealso used a Bayesian-inspired approach for the hybrid methodensemble. This work aimed to test the fundamentalrecommendation techniques on smaller datasets with non-English language using three predictors: user interactions bylike button, click-throughs (and article text itself). Despite therecent popularity of the neural network models, we found thatsimple models outperformed the models with neural networkson our dataset.

Název v anglickém jazyce

Comparison of basic recommendation methods for Czech news articles

Popis výsledku anglicky

The common contemporary recommendationmethods were assessed for recommending relevant news articlesin the Czech language. We used a TF-IDF vectorization withcosine similarity for content-based recommendations andsingular value decomposition for collaborative filtering. Thesimplest model was the popularity model based on the votecounts of other users. While the best days of expert systems seemto be gone in the era of more sophisticated neural networkmodels, we also wanted to investigate whether expert systemscan still show some benefits in providing relatively simple andexplainable methods for an ensemble of recommendations. Wealso used a Bayesian-inspired approach for the hybrid methodensemble. This work aimed to test the fundamentalrecommendation techniques on smaller datasets with non-English language using three predictors: user interactions bylike button, click-throughs (and article text itself). Despite therecent popularity of the neural network models, we found thatsimple models outperformed the models with neural networkson our dataset.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2024

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

2024 IEEE 17th International Scientific Conference on Informatics

ISBN

979-8-3503-8768-1

ISSN

—

e-ISSN

—

Počet stran výsledku

7

Strana od-do

244-250

Název nakladatele

IEEE Institute of Electrical and Electronics Engineers

Místo vydání

New Jersey

Místo konání akce

Poprad

Datum konání akce

13. 11. 2024

Typ akce podle státní příslušnosti

EUR - Evropská akce

Kód UT WoS článku

—