Supplemental Files for the article "Cross-Domain Sentiment Classification in Slovak: A Comparative Analysis of Transformer-Based Models"

The supplementary materials include all code and results relevant to the fine-tuning of sentiment classification models for Slovak text across various domains. Specifically, these files contain:1. Fine-tuning Code for Language Models: Scripts for fine-tuning SlovakBERT, DistilBERT, and SK_BPE_BLM models on Slovak sentiment classification, designed for reproducibility of training and evaluation.2. Training and Evaluation Results: Files detailing each model’s performance across domains (e.g., banking, social media) using 10-fold cross-validation, with F1-scores for sentiment classes.3. Encoding Configurations: Encoding files to ensure consistent text pre-processing and reproducibility of model training.4. Datasets: Domain-specific sentiment-labeled datasets (positive, negative, neutral) offering insight into Slovak language characteristics for NLP research.