arXiv subjects

Klesti Hoxha

Publications and source records attributed to Klesti Hoxha.

3 recordsLinked to original sources

arXiv2026

From Entity Mentions to Tone: An LLM-Based Pipeline for Media Bias Analysis

This paper presents a pipeline for analyzing media bias and framing in online news. The pipeline groups articles into topics and events, adds named-entity and sentiment annotations, and compares news sources through people mentions, source-level tone, and event-level coverage patterns. We apply it to 8,358 Albanian news articles collected from GDELT and compare the resulting annotations with GDELT's automated annotations. The results show moderate agreement for sentiment and entity extraction, as well as additional person-entity pairs that can potentially support the bias analysis. We compare two annotation prompts and find that stricter sentiment-validation rules remove label-score inconsistencies but increase execution time and reduce annotation coverage. Based on these results, the simpler prompt is used for the rest of the analysis. We have provided sample analysis on source-level framing pro les, person-level tone differences across sources, and event-level gatekeeping and coverage indicators. These outputs show how the same news collection can be used to examine what sources cover, how they describe public figures, and where coverage is concentrated. The approach is particularly useful in settings where manually verified datasets or specialized language tools are limited.

cs.CL↗

arXiv2019

Albanian Language Identification in Text Documents

In this work we investigate the accuracy of standard and state-of-the-art language identification methods in identifying Albanian in written text documents. A dataset consisting of news articles written in Albanian has been constructed for this purpose. We noticed a considerable decrease of accuracy when using test documents that miss the Albanian alphabet letters " Ë " and " Ç " and created a custom training corpus that solved this problem by achieving an accuracy of more than 99%. Based on our experiments, the most performing language identification methods for Albanian use a naïve Bayes classifier and n-gram based classification features.

cs.IR↗

arXiv2014

Towards a Modular Recommender System for Research Papers written in Albanian

In the recent years there has been an increase in scientific papers publications in Albania and its neighboring countries that have large communities of Albanian speaking researchers. Many of these papers are written in Albanian. It is a very time consuming task to find papers related to the researchers' work, because there is no concrete system that facilitates this process. In this paper we present the design of a modular intelligent search system for articles written in Albanian. The main part of it is the recommender module that facilitates searching by providing relevant articles to the users (in comparison with a given one). We used a cosine similarity based heuristics that differentiates the importance of term frequencies based on their location in the article. We did not notice big differences on the recommendation results when using different combinations of the importance factors of the keywords, title, abstract and body. We got similar results when using only the title and abstract in comparison with the other combinations. Because we got fairly good results in this initial approach, we believe that similar recommender systems for documents written in Albanian can be build also in contexts not related to scientific publishing.

cs.IR↗