File: version-2-5.text

package info (click to toggle)
autoclass 3.3.4-6
  • links: PTS
  • area: main
  • in suites: lenny
  • size: 3,844 kB
  • ctags: 994
  • sloc: ansic: 16,674; makefile: 123; sh: 98; cpp: 95; csh: 77
file content (98 lines) | stat: -rw-r--r-- 3,566 bytes parent folder | download | duplicates (7)
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
91
92
93
94
95
96
97
98


	AUTOCLASS C VERSION 2.5 NOTES 
====================================================================== 
======================================================================

Documentation:
------------------------------

  1. autoclass-c/doc/reports-c.text -
        Minor typographical changes.  Added new report generation
        parameter: order_attributes_by_influence_p.  Its default
        value is true.  The file extension of the influence values
        report has been changed from ".influ-text-1" to ".influ-o-text-1"
        when order_attributes_by_influence_p = true, and to
        ".influ-no-text-1" when order_attributes_by_influence_p = false.

  2. autoclass-c/doc/interpretation-c.text -
        Minor changes to the text.

  3. autoclass-c/sample/imports-85c.influ-o-text-1
        Influence values report has been significantly revised and
        reformatted.

  4. autoclass-c/doc/search-c.text -
        Corrected definition of fixed_j.


Programming:
------------------------------						

  1. autoclass-c/prog/globals.c -
        Update "G_ac_version" to 2.5.

  2. autoclass-c/prog/intf-reports.c, utils.c -
        Formatting change to "format_real_attribute" for multiple
        multivariate attribute groups.

        Remove covariance matrix output and reformat the correlation
        matrix output to fixed decimal point notation.

        For the influence values report, sort real valued attributes
        of the same model group by the first significance value, if
        that group is multi_normal_cn.

        For discrete attributes: relabel the headers "Prob", rather than
        "Mean"; and correct the instance value significance computation
        to be "local_prob * log( local_prob / global_prob)".

  3. autoclass-c/prog/autoclass.h -
        Add #ifndef for MAXPATHLEN.

  4. autoclass-c/prog/io-results.c -
        In "validate_data_pathname", "validate_results_pathname", &
        "make_and_validate_pathname", only do fclose, if fopen returns
        non-NULL.

  5. autoclass-c/prog/search-control-2.c -
        Add "pad" argument to "print_search_try".

  6. autoclass-c/prog/intf-extensions.c -
        Formatting change to "get_models_source_info".

  7. autoclass-c/load-ac,
     autoclass-c/prog/autoclass.make.sun.gcc, autoclass.make.sun.acc,
     autoclass.make.sun.cc, autoclass.make.sgi -
     (remove autoclass.make.sun)
        Add SunOS/Solaris C compiler support.

  8. autoclass-c/prog/io-results.c, io-read-model.c, io-read-data.c,
     utils.c, intf-reports.c, getparmas.c -
        Cast return values of "strlen" to int.

  9. autoclass-c/prog/model-transforms.c -
        In "generate_singleton_transform", correct segmentation fault
        which occurs when more than 25 type = real, subtype = scalar
        attributes are defined in the ".hd2" & ".model" files.

 10. autoclass-c/prog/struct-data.c, io-results.c, io-results-bin.c -
        Properly initialize att_info array when it exceeds preallocated
        size. 

 11. autoclass-c/load-ac,
     autoclass-c/prog/autoclass.make.linux.gcc, autoclass.make.sun.*,
     autoclass.make.sgi, fcntlcom-ac.h -
        Thanks to Andrew Lewycky <plewycky@oise.on.ca>, added mods for
        port to Linux version 1.2.10, GCC version 2.5.8, libc version
        4.6.25.

 12. autoclass-c/prog/model-single-multitnomial.c -
        In "sm_params_influence_fn", add check for out-of-bounds 
        arguments to the log function to prevent "log domain" errors.


======================================================================