-
Notifications
You must be signed in to change notification settings - Fork 1
/
page-blocks.names
91 lines (72 loc) · 3.81 KB
/
page-blocks.names
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
1. Title of Database: Blocks Classification
2. Sources:
(a) Donato Malerba
Dipartimento di Informatica
University of Bari
via Orabona 4
70126 Bari - Italy
phone: +39 - 80 - 5443269
fax: +39 - 80 - 5443196
(b) Donor: Donato Malerba
(c) Date: July 1995
3. Past Usage:
This data set have been used to try different simplification methods
for decision trees. A summary of the results can be found in:
Malerba, D., Esposito, F., and Semeraro, G.
"A Further Comparison of Simplification Methods for Decision-Tree Induction."
In D. Fisher and H. Lenz (Eds.), "Learning from Data:
Artificial Intelligence and Statistics V", Lecture Notes in Statistics,
Springer Verlag, Berlin, 1995.
The problem consists in classifying all the blocks of the page
layout of a document that has been detected by a segmentation
process. This is an essential step in document analysis
in order to separate text from graphic areas. Indeed,
the five classes are: text (1), horizontal line (2),
picture (3), vertical line (4) and graphic (5).
For a detailed presentation of the problem see:
Esposito F., Malerba D., & Semeraro G.
Multistrategy Learning for Document Recognition
Applied Artificial Intelligence, 8, pp. 33-84, 1994
All instances have been personally checked so that
low noise is present in the data.
4. Relevant Information Paragraph:
The 5473 examples comes from 54 distinct documents.
Each observation concerns one block.
All attributes are numeric.
Data are in a format readable by C4.5.
5. Number of Instances: 5473.
6. Number of Attributes
height: integer. | Height of the block.
lenght: integer. | Length of the block.
area: integer. | Area of the block (height * lenght);
eccen: continuous. | Eccentricity of the block (lenght / height);
p_black: continuous. | Percentage of black pixels within the block (blackpix / area);
p_and: continuous. | Percentage of black pixels after the application of the Run Length Smoothing Algorithm (RLSA) (blackand / area);
mean_tr: continuous. | Mean number of white-black transitions (blackpix / wb_trans);
blackpix: integer. | Total number of black pixels in the original bitmap of the block.
blackand: integer. | Total number of black pixels in the bitmap of the block after the RLSA.
wb_trans: integer. | Number of white-black transitions in the original bitmap of the block.
7. Missing Attribute Values: No missing value.
8. Class Distribution:
Valid Cum
Class Frequency Percent Percent Percent
text 4913 89.8 89.8 89.8
horiz. line 329 6.0 6.0 95.8
graphic 28 .5 .5 96.3
vert. line 88 1.6 1.6 97.9
picture 115 2.1 2.1 100.0
------- ------- -------
TOTAL 5473 100.0 100.0
Summary Statistics:
Variable Mean Std Dev Minimum Maximum Correlation
HEIGHT 10.47 18.96 1 804 .3510
LENGTH 89.57 114.72 1 553 -.0045
AREA 1198.41 4849.38 7 143993 .2343
ECCEN 13.75 30.70 .007 537.00 .0992
P_BLACK .37 .18 .052 1.00 .2130
P_AND .79 .17 .062 1.00 -.1771
MEAN_TR 6.22 69.08 1.00 4955.00 .0723
BLACKPIX 365.93 1270.33 7 33017 .1656
BLACKAND 741.11 1881.50 7 46133 .1565
WB_TRANS 106.66 167.31 1 3212 .0337