anneal.arff

是UCI数据库中的一些有代表性的数据集

% 1. Title of Database: Annealing Data
% 
% 2. Source Information: donated by David Sterling and Wray Buntine.
% 
% 3. Past Usage: unknown
% 
% 4. Relevant Information:
%    -- Explanation: I suspect this was left by Ross Quinlan in 1987 at the
%       4th Machine Learning Workshop.  I'd have to check with Jeff Schlimmer
%       to double check this.
% 
% 5. Number of Instances: 798
% 
% 6. Number of Attributes: 38
%    -- 6 continuously-valued
%    -- 3 integer-valued
%    -- 29 nominal-valued
% 
% 7. Attribute Information:
%     1. family:          --,GB,GK,GS,TN,ZA,ZF,ZH,ZM,ZS
%     2. product-type:    C, H, G
%     3. steel:           -,R,A,U,K,M,S,W,V
%     4. carbon:          continuous
%     5. hardness:        continuous
%     6. temper_rolling:  -,T
%     7. condition:       -,S,A,X
%     8. formability:     -,1,2,3,4,5
%     9. strength:        continuous
%    10. non-ageing:      -,N
%    11. surface-finish:  P,M,-
%    12. surface-quality: -,D,E,F,G
%    13. enamelability:   -,1,2,3,4,5
%    14. bc:              Y,-
%    15. bf:              Y,-
%    16. bt:              Y,-
%    17. bw/me:           B,M,-
%    18. bl:              Y,-
%    19. m:               Y,-
%    20. chrom:           C,-
%    21. phos:            P,-
%    22. cbond:           Y,-
%    23. marvi:           Y,-
%    24. exptl:           Y,-
%    25. ferro:           Y,-
%    26. corr:            Y,-
%    27. blue/bright/varn/clean:          B,R,V,C,-
%    28. lustre:          Y,-
%    29. jurofm:          Y,-
%    30. s:               Y,-
%    31. p:               Y,-
%    32. shape:           COIL, SHEET
%    33. thick:           continuous
%    34. width:           continuous
%    35. len:             continuous
%    36. oil:             -,Y,N
%    37. bore:            0000,0500,0600,0760
%    38. packing: -,1,2,3
%    classes:        1,2,3,4,5,U
%  
%    -- The '-' values are actually 'not_applicable' values rather than
%       'missing_values' (and so can be treated as legal discrete
%       values rather than as showing the absence of a discrete value).
% 
% 8. Missing Attribute Values: Signified with "?"
%    Attribute:  Number of instances missing its value:
%    1           0
%    2           0
%    3           70
%    4           0
%    5           0
%    6           675
%    7           271
%    8           283
%    9           0
%   10           703
%   11           790
%   12           217
%   13           785
%   14           797
%   15           680
%   16           736
%   17           609
%   18           662
%   19           798
%   20           775
%   21           791
%   22           730
%   23           798
%   24           796
%   25           772
%   26           798
%   27           793
%   28           753
%   29           798
%   30           798
%   31           798
%   32           0
%   33           0
%   34           0
%   35           0
%   36           740
%   37           0
%   38           789
%   39           0
% 
% 9. Distribution of Classes
%      Class Name:   Number of Instances:
%      1               8
%      2              88
%      3             608
%      4               0