----- What: ----- This is the readme file for the causal structure discovery weighted maxSAT data sets. These data sets are part of publication: A. Hyttinen, P. Saikko, M. Järvisalo: A Core-Guided Approach to Learning Optimal Causal Graphs International Joint Conference in Artificial Intelligence, IJCAI 2017. ---------- References ---------- The data sets are the product of the following original research papers: A. Hyttinen, P. Saikko, M. Järvisalo: A Core-Guided Approach to Learning Optimal Causal Graphs International Joint Conference in Artificial Intelligence, IJCAI 2017. J. Berg, A. Hyttinen, M. Järvisalo: "Applications of MaxSAT in Data Analysis" Pragmatics of SAT workshop 2015. A. Hyttinen, F. Eberhardt, and M. Järvisalo: "Constraint-based Causal Discovery: Conflict Resolution with Answer Set Programming", Proceedings of the 30th Conference on Uncertainty in Artificial Intelligence, 2014. A. Hyttinen, P. O. Hoyer, F. Eberhardt, and M. Järvisalo: "Discovering Cyclic Causal Models with Latent Variables: A General SAT-Based Procedure", Proceedings of the 29th Conference on Uncertainty in Artificial Intelligence, 2013. Also presented at Pragmatics of SAT workshop 2013 and Approaches to Causal Structure Learning Workshop 2013. ----------------------- SELECTION OF INSTANCES ----------------------- The benchmarks are from real-world datasets often used for benchmarking exact Bayesian network structure learning algorithms. We considered suitable-sized subsets of the variables in the datasets, the remaining variables becoming thus latent (causal graph definition supports latent variables). We employed the BDEU score with equivalent sample size 10 to obtain independence constraint weights for this discrete data. The were turned to intergers by multiplying by 1000 and rounding. All files use the encoding over conditioning and marginalization operations (Hyttinen et al. 2014). The number of variables was selected for each data such that we would get a sensible and fair comparison among the solvers studied. ----------- FILE NAMES ----------- Files are named following convention: causal___.wcnf where = the dataset name from which the instance was generated from. = number of observed variables in the causal graph = the number of samples ------- CONTACT ------- In case of questions please check the original papers first, then you can contact: Antti Hyttinen email: antti.hyttinen@helsinki.fi --------- DATASETS: --------- 1. causal_Adult_6_30162.wcnf 2. causal_Alarm_7_1000.wcnf 3. causal_alarm_7_100.wcnf 4. causal_alarm_9_10000.wcnf 5. causal_alarm_9_1000.wcnf 6. causal_asia_10_100.wcnf 7. causal_asia_7_1000.wcnf 8. causal_asia_8_10000.wcnf 9. causal_Autos_8_159.wcnf 10. causal_Bands_6_277.wcnf 11. causal_carpo_8_1000.wcnf 12. causal_carpo_8_100.wcnf 13. causal_carpo_9_10000.wcnf 14. causal_Diabetes_8_10000.wcnf 15. causal_Diabetes_8_1000.wcnf 16. causal_Diabetes_8_100.wcnf 17. causal_Epigenetics_7_72228.wcnf 18. causal_Flag_10_194.wcnf 19. causal_hailfinder_7_100.wcnf 20. causal_hailfinder_8_10000.wcnf 21. causal_hailfinder_9_1000.wcnf 22. causal_Heart_10_212.wcnf 23. causal_Hepatitis_10_126.wcnf 24. causal_Horse.23_7_300.wcnf 25. causal_Horse_9_300.wcnf 26. causal_Image_7_2310.wcnf 27. causal_Imports_6_205.wcnf 28. causal_insurance_8_10000.wcnf 29. causal_insurance_9_100.wcnf 30. causal_Letter_7_20000.wcnf 31. causal_Link_10_10000.wcnf 32. causal_Link_10_1000.wcnf 33. causal_Link_9_100.wcnf 34. causal_LungCancer_8_27.wcnf 35. causal_Meta_7_528.wcnf 36. causal_Mildew_6_100.wcnf 37. causal_Mildew_8_1000.wcnf 38. causal_Mildew_9_10000.wcnf 39. causal_Mushroom_7_1000.wcnf 40. causal_Mushroom_7_8124.wcnf 41. causal_Parkinsons_6_195.wcnf 42. causal_Pigs_6_10000.wcnf 43. causal_Pigs_6_1000.wcnf 44. causal_Pigs_6_100.wcnf 45. causal_Sensors_7_5456.wcnf 46. causal_Soybean_9_266.wcnf 47. causal_Spectf_10_267.wcnf 48. causal_Statlog_7_752.wcnf 49. causal_SteelPlates_6_1941.wcnf 50. causal_Voting_7_435.wcnf 51. causal_Water_10_1000.wcnf 52. causal_Water_10_100.wcnf 53. causal_Water_7_10000.wcnf 54. causal_Water_7_380.wcnf 55. causal_Wdbc_8_569.wcnf 56. causal_Wine_8_178.wcnf 57. causal_Zoo_6_101.wcnf