source: orange/orange/doc/datasets/lymphography.htm @ 1760:9d4bb141fb0e

Revision 1760:9d4bb141fb0e, 3.3 KB checked in by blaz <blaz.zupan@…>, 9 years ago (diff)

data info file

Line 
1<html>
2<head>
3<title>Lymphography Data Base</title>
4</head>
5<body>
6<h1>Info on Lymphography Data Base</h1>
7<pre>
8Citation Request:
9   This lymphography domain was obtained from the University Medical Centre,
10   Institute of Oncology, Ljubljana, Yugoslavia.  Thanks go to M. Zwitter and
11   M. Soklic for providing the data.  Please include this citation if you plan
12   to use this database.
13
141. Title: Lymphography Domain
15
162. Sources:
17    (a) See Above.
18    (b) Donors: Igor Kononenko,
19                University E.Kardelj
20                Faculty for electrical engineering
21                Trzaska 25
22                61000 Ljubljana (tel.: (38)(+61) 265-161
23
24                Bojan Cestnik
25                Jozef Stefan Institute
26                Jamova 39
27                61000 Ljubljana
28                Yugoslavia (tel.: (38)(+61) 214-399 ext.287)
29   (c) Date: November 1988
30
313. Past Usage: (sveral)
32    1. Cestnik,G., Konenenko,I, & Bratko,I. (1987). Assistant-86: A
33       Knowledge-Elicitation Tool for Sophisticated Users.  In I.Bratko
34       & N.Lavrac (Eds.) Progress in Machine Learning, 31-45, Sigma Press.
35       -- Assistant-86: 76% accuracy
36    2. Clark,P. & Niblett,T. (1987). Induction in Noisy Domains.  In
37       I.Bratko & N.Lavrac (Eds.) Progress in Machine Learning, 11-30,
38       Sigma Press.
39       -- Simple Bayes: 83% accuracy
40       -- CN2 (99% threshold): 82%
41    3. Michalski,R., Mozetic,I. Hong,J., & Lavrac,N. (1986).  The Multi-Purpose
42       Incremental Learning System AQ15 and its Testing Applications to Three
43       Medical Domains.  In Proceedings of the Fifth National Conference on
44       Artificial Intelligence, 1041-1045. Philadelphia, PA: Morgan Kaufmann.
45       -- Experts: 85% accuracy (estimate)
46       -- AQ15: 80-82%
47
484. Relevant Information:
49     This is one of three domains provided by the Oncology Institute
50     that has repeatedly appeared in the machine learning literature.
51     (See also breast-cancer and primary-tumor.)
52
535. Number of Instances: 148
54
556. Number of Attributes: 19 including the class attribute
56
577. Attribute information:
58    --- NOTE: All attribute values in the database have been entered as
59              numeric values corresponding to their index in the list
60              of attribute values for that attribute domain as given below.
61    1. class: normal find, metastases, malign lymph, fibrosis
62    2. lymphatics: normal, arched, deformed, displaced
63    3. block of affere: no, yes
64    4. bl. of lymph. c: no, yes
65    5. bl. of lymph. s: no, yes
66    6. by pass: no, yes
67    7. extravasates: no, yes
68    8. regeneration of: no, yes
69    9. early uptake in: no, yes
70   10. lym.nodes dimin: 0-3
71   11. lym.nodes enlar: 1-4
72   12. changes in lym.: bean, oval, round
73   13. defect in node: no, lacunar, lac. marginal, lac. central
74   14. changes in node: no, lacunar, lac. margin, lac. central
75   15. changes in stru: no, grainy, drop-like, coarse, diluted, reticular,
76                        stripped, faint,
77   16. special forms: no, chalices, vesicles
78   17. dislocation of: no, yes
79   18. exclusion of no: no, yes
80   19. no. of nodes in: 0-9, 10-19, 20-29, 30-39, 40-49, 50-59, 60-69, >=70
81
828. Missing Attribute Values: None
83
849. Class Distribution:
85    Class:        Number of Instances:
86    normal find:  2
87    metastases:   81
88    malign lymph: 61
89    fibrosis:     4
90</pre>
91</body>
92</html>
Note: See TracBrowser for help on using the repository browser.