18. c6: {x,o,b}
19. d1: {x,o,b}
20. d2: {x,o,b}
21. d3: {x,o,b}
22. d4: {x,o,b}
23. d5: {x,o,b}
24. d6: {x,o,b}
25. e1: {x,o,b}
26. e2: {x,o,b}
27. e3: {x,o,b}
28. e4: {x,o,b}
29. e5: {x,o,b}
30. e6: {x,o,b}
31. f1: {x,o,b}
32. f2: {x,o,b}
33. f3: {x,o,b}
34. f4: {x,o,b}
35. f5: {x,o,b}
36. f6: {x,o,b}
37. g1: {x,o,b}
38. g2: {x,o,b}
39. g3: {x,o,b}
40. g4: {x,o,b}
41. g5: {x,o,b}
42. g6: {x,o,b}
43. Class: {win,loss,draw}
"""
if __name__ == "__main__":
util.check_dict(convert())
to 13 contain static features and columns 14 to 51 dynamic
features.
7. Description of attributes:
The full names for each attribute are provided in the paper, tables
2 and 3.
Raw data: all attributes are numeric. Attributes 5, 9, 11, 13
and 35 are integer-valued. All other attributes are continuous.
Training, validation and test data: all data are numeric and
continuous on account of being normalized.
8. Missing Attribute Values:
There are no missing values.
9. Class Distribution: number of positive instances in the sets for
each heuristic (H1 to H5) and the "decline" option H0.
H1 H2 H3 H4 H5 H0
Training set: 556 229 373 303 312 1286
Validation set: 260 133 187 146 159 644
Test set: 273 124 188 168 153 624
"""
if __name__ == "__main__":
util.check_dict(convert())
T20: continuous.
T21: continuous.
T22: continuous.
T23: continuous.
T_PK: continuous.
T_AV: continuous.
T85: continuous.
RH85: continuous.
U85: continuous.
V85: continuous.
HT85: continuous.
T70: continuous.
RH70: continuous.
U70: continuous.
V70: continuous.
HT70: continuous.
T50: continuous.
RH50: continuous.
U50: continuous.
V50: continuous.
HT50: continuous.
KI: continuous.
TT: continuous.
SLP: continuous.
SLP_: continuous.
Precp: continuous.
"""
if __name__ == "__main__":
check_dict(convert())
