Replication Package for: Dissecting NetDoktor With Entropy, Tree Simplification & Logical Abduction

Replication package for: "Dissecting NetDoktor’s Online Questionnaire With Entropy, Tree Simplification & Logical Abduction". This package includes data and code to replicate and visualize the main results of the paper. The dataset used in this work was scraped from NetDoktor[1], which we explained in our previous work[2,3]. We name this dataset SCQ and produced three structural variants: SCQ_S.jsonl (simplified/pruned), SCQ_E.jsonl (re-constructed based on entropy), and SCQ_balanced.jsonl. While we do include all structural variants, including the original, we only include ID-referenced entries representing the original texts. This is done for two reasons: 1) We do not hold the copyright, and 2) we do not want to leak the data, which would compromise its future use for LLM testing. The system descriptions SD_SCQ.lp, SD_SCQ_S.lp, and SD_SCQ_E.lp are derived from the structural variants using the algorithm described in the paper and the confidence thresholds τ⁺ = 0 (all "Ja" answers are represented as rules), τ⁻ = 0.8, where "Ja" = 7 and "Nein" = 8 in the data source files. [1] Richter et al. (accessed 05.05.2026). Symptom-Checker. Online: www.netdoktor.at/symptom-checker[2] Perko and Wotawa (2024). Testing ChatGPT's Performance on Medical Diagnostic Tasks. In Proceedings of the 27th International Multiconference Information Society – IS 2024. Information Society.[3] Mujić et al. (2025). Extraction of Knowledge Representations for Reasoning from Medical Questionnaires. In Proceedings of the 28th International Multiconference Information Society – IS 2025. Information Society.