search

ai-se / george

stuff for george
MIT License
1 stars 0 forks source link

Kloc between Kloc*(1 +/- noise) where noise in [0.2,0.4,0.6,0.8,1.0] #23

Open bigfatnoob opened 8 years ago

bigfatnoob commented 8 years ago

Mystery1

rank ,                   name ,    med   ,  iqr 
----------------------------------------------------
1 ,           COCOMO2:0.4n ,   31 ,   31 (   --*--       |              ), 0.22,  0.31,  0.53
1 ,           COCOMO2:1.0n ,   37 ,   31 (   ---*--      |              ), 0.24,  0.37,  0.55
1 ,           COCOMO2:0.2n ,   41 ,   44 (   ---*---     |              ), 0.22,  0.41,  0.66
1 ,                COCOMO2 ,   43 ,   35 (    ---*-      |              ), 0.25,  0.43,  0.60
1 ,           COCOMO2:0.6n ,   49 ,   54 (  ------*--    |              ), 0.14,  0.49,  0.68
2 ,           COCOMO2:0.8n ,   60 ,   33 (        -*---  |              ), 0.50,  0.60,  0.83

Mystery2

rank ,                   name ,    med   ,  iqr 
----------------------------------------------------
1 ,           COCOMO2:0.2n ,   43 ,  141 (*              |              ), 0.18,  0.43,  1.59
1 ,           COCOMO2:0.4n ,   45 ,  137 (*              |              ), 0.22,  0.45,  1.59
1 ,                COCOMO2 ,   46 ,  134 (*              |              ), 0.20,  0.46,  1.54
1 ,           COCOMO2:1.0n ,   49 ,  114 (*              |              ), 0.23,  0.49,  1.37
1 ,           COCOMO2:0.6n ,   57 ,  104 (*              |              ), 0.28,  0.57,  1.32
1 ,           COCOMO2:0.8n ,   58 ,  122 (*              |              ), 0.34,  0.58,  1.56

nasa93

rank ,                   name ,    med   ,  iqr 
----------------------------------------------------
1 ,           COCOMO2:1.0n ,   38 ,   38 (-*             |              ), 0.19,  0.38,  0.57
1 ,                COCOMO2 ,   39 ,   39 (-*             |              ), 0.21,  0.39,  0.60
1 ,           COCOMO2:0.2n ,   40 ,   34 (-*             |              ), 0.22,  0.40,  0.56
1 ,           COCOMO2:0.4n ,   41 ,   33 (-*             |              ), 0.27,  0.41,  0.60
2 ,           COCOMO2:0.6n ,   48 ,   43 (-*             |              ), 0.26,  0.48,  0.69
2 ,           COCOMO2:0.8n ,   57 ,   39 ( *             |              ), 0.34,  0.57,  0.73

coc81

rank ,                   name ,    med   ,  iqr 
----------------------------------------------------
1 ,                COCOMO2 ,   32 ,   33 ( --*           |              ), 0.21,  0.32,  0.54
1 ,           COCOMO2:1.0n ,   35 ,   32 ( --*           |              ), 0.21,  0.35,  0.53
1 ,           COCOMO2:0.2n ,   36 ,   29 (  -*           |              ), 0.22,  0.36,  0.51
2 ,           COCOMO2:0.6n ,   44 ,   41 (  --*          |              ), 0.24,  0.44,  0.65
2 ,           COCOMO2:0.4n ,   45 ,   38 (  --*          |              ), 0.22,  0.45,  0.60
2 ,           COCOMO2:0.8n ,   51 ,   53 ( ---*-         |              ), 0.20,  0.51,  0.73
timm commented 8 years ago

now is 1.0 up to 200% wrong?

and am i reading this right? there is no trend of increased error ==> lower ranks. we over get 1.0 and 0.2 in rank1 and 0.4,0.6,0.8 in rank2?

bigfatnoob commented 8 years ago

now is 1.0 up to 200% wrong?

Yes 1.0 Implies KLOC varies between 0 and 2 times KLOC

and am i reading this right? there is no trend of increased error ==> lower ranks. we over get 1.0 and 0.2 in rank1 and 0.4,0.6,0.8 in rank2?

I think the way the experiment is conducted might not be right. For example at 0.2 we select KLOC randomly between randomly between 0.8KLOC and 1.2KLOC. For 0.4 we select KLOC randomly between 0.6KLOC and 1.4KLOC. Thus the range of 0.2 lies within the range of 0.4. Moreover this experiment is performed only once for each project so the effect might be just because of randomness. I will repeat it 20 times

timm commented 8 years ago

huh. i hear u, so not 0.8 to 1.2 but

0.25 to 1 0.5 to 1 0.75 to 1 1 to 1.25 1 to 1.5 1 to 1.75 1 to 2

repeated 20 times?

bigfatnoob commented 8 years ago

Mystery1

Mean

rank ,                   name ,    med   ,  iqr 
----------------------------------------------------
1 ,           COCOMO2:0.2n ,   38 ,   32 (    ---*----   |              ), 0.25,  0.38,  0.57
1 ,                COCOMO2 ,   43 ,   35 (    ----*---   |              ), 0.25,  0.43,  0.60
1 ,           COCOMO2:0.4n ,   45 ,   40 (     ----*-----|              ), 0.30,  0.45,  0.70
2 ,           COCOMO2:0.6n ,   49 ,   30 (      ----*--  |              ), 0.32,  0.49,  0.62
2 ,           COCOMO2:0.8n ,   49 ,   36 (       ---*----|              ), 0.39,  0.49,  0.75
2 ,           COCOMO2:1.0n ,   50 ,   34 (       ---*----|              ), 0.39,  0.50,  0.73

Variance

rank ,                   name ,    med   ,  iqr 
----------------------------------------------------
1 ,                COCOMO2 ,    0 ,    0 (*              |              ), 0.00,  0.00,  0.00
1 ,           COCOMO2:0.2n ,    1 ,    1 (*              |              ), 0.00,  0.01,  0.01
2 ,           COCOMO2:0.4n ,    2 ,    2 (*              |              ), 0.02,  0.02,  0.04
2 ,           COCOMO2:0.6n ,    4 ,    2 (-*             |              ), 0.03,  0.04,  0.05
3 ,           COCOMO2:0.8n ,    6 ,    3 ( *             |              ), 0.05,  0.06,  0.08
3 ,           COCOMO2:1.0n ,    8 ,    7 ( -*            |              ), 0.05,  0.08,  0.12

Mystery2

Mean

rank ,                   name ,    med   ,  iqr 
----------------------------------------------------
1 ,           COCOMO2:0.2n ,   46 ,  132 (*              |              ), 0.19,  0.46,  1.51
1 ,                COCOMO2 ,   46 ,  134 (*              |              ), 0.20,  0.46,  1.54
1 ,           COCOMO2:0.4n ,   48 ,  131 (*              |              ), 0.25,  0.48,  1.56
1 ,           COCOMO2:0.6n ,   50 ,  120 (*              |              ), 0.36,  0.50,  1.56
1 ,           COCOMO2:0.8n ,   63 ,  100 (*              |              ), 0.46,  0.63,  1.46
1 ,           COCOMO2:1.0n ,   65 ,  113 (*              |              ), 0.51,  0.65,  1.64

Variance

rank ,                   name ,    med   ,  iqr 
----------------------------------------------------
1 ,                COCOMO2 ,    0 ,    0 (*              |              ), 0.00,  0.00,  0.00
2 ,           COCOMO2:0.2n ,    2 ,    9 (*              |              ), 0.01,  0.02,  0.10
3 ,           COCOMO2:0.4n ,    5 ,   36 (*              |              ), 0.02,  0.05,  0.38
3 ,           COCOMO2:0.6n ,    8 ,   80 (*              |              ), 0.04,  0.08,  0.84
4 ,           COCOMO2:0.8n ,   12 ,  130 (*              |              ), 0.06,  0.12,  1.36
4 ,           COCOMO2:1.0n ,   18 ,  162 (*              |              ), 0.08,  0.18,  1.70

nasa93

Mean

rank ,                   name ,    med   ,  iqr 
----------------------------------------------------
1 ,           COCOMO2:0.2n ,   39 ,   36 (-*             |              ), 0.22,  0.39,  0.58
1 ,                COCOMO2 ,   39 ,   39 (-*             |              ), 0.21,  0.39,  0.60
1 ,           COCOMO2:0.4n ,   40 ,   33 ( *             |              ), 0.24,  0.40,  0.57
1 ,           COCOMO2:0.6n ,   42 ,   26 ( *             |              ), 0.30,  0.42,  0.56
2 ,           COCOMO2:0.8n ,   44 ,   23 ( -*            |              ), 0.37,  0.44,  0.60
2 ,           COCOMO2:1.0n ,   51 ,   18 (  *            |              ), 0.44,  0.51,  0.62

Variance

rank ,                   name ,    med   ,  iqr 
----------------------------------------------------
1 ,                COCOMO2 ,    0 ,    0 (*              |              ), 0.00,  0.00,  0.00
1 ,           COCOMO2:0.2n ,    1 ,    1 (*              |              ), 0.00,  0.01,  0.01
2 ,           COCOMO2:0.4n ,    2 ,    1 (*              |              ), 0.01,  0.02,  0.02
3 ,           COCOMO2:0.6n ,    3 ,    3 (*              |              ), 0.02,  0.03,  0.05
4 ,           COCOMO2:0.8n ,    6 ,    3 (*              |              ), 0.04,  0.06,  0.07
4 ,           COCOMO2:1.0n ,    8 ,    3 (*              |              ), 0.06,  0.08,  0.09

coc81

Mean

rank ,                   name ,    med   ,  iqr 
----------------------------------------------------
1 ,           COCOMO2:0.2n ,   32 ,   32 (  --*-         |              ), 0.22,  0.32,  0.54
1 ,                COCOMO2 ,   32 ,   33 (  --*-         |              ), 0.21,  0.32,  0.54
1 ,           COCOMO2:0.4n ,   36 ,   28 (   -*-         |              ), 0.24,  0.36,  0.52
2 ,           COCOMO2:0.6n ,   43 ,   15 (    -*         |              ), 0.37,  0.43,  0.52
3 ,           COCOMO2:0.8n ,   49 ,   23 (     -*-       |              ), 0.39,  0.49,  0.62
3 ,           COCOMO2:1.0n ,   50 ,   19 (     -*-       |              ), 0.45,  0.50,  0.64

Variance

rank ,                   name ,    med   ,  iqr 
----------------------------------------------------
1 ,                COCOMO2 ,    0 ,    0 (*              |              ), 0.00,  0.00,  0.00
1 ,           COCOMO2:0.2n ,    1 ,    1 (*              |              ), 0.00,  0.01,  0.01
2 ,           COCOMO2:0.4n ,    2 ,    3 (*              |              ), 0.01,  0.02,  0.04
3 ,           COCOMO2:0.6n ,    4 ,    3 (*              |              ), 0.03,  0.04,  0.06
3 ,           COCOMO2:0.8n ,    6 ,    5 (-*             |              ), 0.05,  0.06,  0.10
4 ,           COCOMO2:1.0n ,    9 ,    5 ( *             |              ), 0.07,  0.09,  0.12
bigfatnoob commented 8 years ago

Latex Result Here:

https://www.sharelatex.com/project/5730dc610f37a50b1532a8a0

timm commented 8 years ago

60 == 0.4 to 1.6

note that 20 = 0.8 to 1.2 works just fine