Skip to content
GitLab
Explore
Sign in
Primary navigation
Search or go to…
Project
T
TP Optimisation
Manage
Activity
Members
Labels
Plan
Issues
Issue boards
Milestones
Wiki
Code
Merge requests
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Snippets
Build
Pipelines
Jobs
Pipeline schedules
Artifacts
Deploy
Releases
Package registry
Model registry
Operate
Environments
Terraform modules
Monitor
Incidents
Analyze
Value stream analytics
Contributor analytics
CI/CD analytics
Repository analytics
Model experiments
Help
Help
Support
GitLab documentation
Compare GitLab plans
GitLab community forum
Contribute to GitLab
Provide feedback
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
stefano.cirieco
TP Optimisation
Commits
80c0c057
Commit
80c0c057
authored
3 years ago
by
stefano.cirieco
Browse files
Options
Downloads
Patches
Plain Diff
Mis à jour pour la validation croisée
parent
66392c76
No related branches found
No related tags found
No related merge requests found
Changes
4
Show whitespace changes
Inline
Side-by-side
Showing
4 changed files
graph.py
+25
-20
25 additions, 20 deletions
graph.py
main.c
+76
-14
76 additions, 14 deletions
main.c
opti_num.c
+27
-10
27 additions, 10 deletions
opti_num.c
opti_num.h
+2
-2
2 additions, 2 deletions
opti_num.h
with
130 additions
and
46 deletions
graph.py
+
25
−
20
View file @
80c0c057
...
...
@@ -2,29 +2,34 @@ import sys
import
matplotlib.pyplot
as
plt
import
numpy
as
np
x
=
[]
y
=
[]
a
=
0
b
=
0
x
=
[
[],
[],
[]
]
y
=
[
[],
[],
[]
]
a
=
[]
b
=
[]
with
open
(
'
points.txt
'
)
as
f
:
lines
=
f
.
readlines
(
)
fig
,
axes
=
plt
.
subplots
(
3
,
1
,
figsize
=
(
18
,
15
))
plt
.
get_current_fig_manager
().
set_window_title
(
'
Graphics
'
)
z
=
lines
[
0
].
split
(
"
,
"
)
a
=
float
(
z
[
0
])
b
=
float
(
z
[
1
])
for
i
in
range
(
3
):
with
open
(
'
G
'
+
str
(
i
+
1
)
+
'
.txt
'
)
as
f
:
for
line
in
f
.
readlines
():
z
=
line
.
split
(
"
,
"
)
x
[
i
].
append
(
float
(
z
[
0
]))
y
[
i
].
append
(
float
(
z
[
1
]))
for
line
in
lines
[
1
:]:
with
open
(
'
ab.txt
'
)
as
f
:
for
line
in
f
.
readlines
():
z
=
line
.
split
(
"
,
"
)
x
.
append
(
float
(
z
[
0
]))
y
.
append
(
float
(
z
[
1
]))
a
.
append
(
float
(
z
[
0
]))
b
.
append
(
float
(
z
[
1
]))
fig
,
ax
=
plt
.
subplots
()
titles
=
[(
2
,
r
'
$G1 \bigcup G2 \longmapsto G3$
'
),
(
0
,
r
'
$G2 \bigcup G3 \longmapsto G1$
'
),
(
1
,
r
'
$G1 \bigcup G3 \longmapsto G2$
'
)]
# 1st point (0, b) and 2nd point (b, a+b)
xLine
,
yLine
=
[
0
,
1
],
[
b
,
a
+
b
]
plt
.
plot
(
xLine
,
yLine
,
marker
=
'
o
'
)
ax
.
plot
(
x
,
y
,
'
o
'
,
color
=
'
tab:brown
'
)
for
i
in
range
(
3
):
axes
[
i
].
set_title
(
titles
[
i
][
1
])
axes
[
i
].
axline
([
0
,
b
[
i
]],
[
1
,
a
[
i
]
+
b
[
i
]])
axes
[
i
].
plot
(
x
[
titles
[
i
][
0
]],
y
[
titles
[
i
][
0
]],
'
.
'
,
color
=
'
tab:green
'
)
axes
[
i
].
grid
(
True
)
plt
.
show
()
\ No newline at end of file
This diff is collapsed.
Click to expand it.
main.c
+
76
−
14
View file @
80c0c057
...
...
@@ -2,18 +2,21 @@
#include
<stdlib.h>
#include
<math.h>
#include
<time.h>
//#include <system.h>
#include
"vec.h"
#include
"opti_num.h"
#define NB_G 3
double
randNum
();
void
write_to_file
(
vec
ab
,
vec
*
points
,
int
N
,
char
*
filename
);
int
*
rand_order_indexes
(
int
N
);
vec
**
cut_table
(
int
*
rand_indexes
,
int
N
,
vec
*
points
);
void
write_to_file
(
vec
*
points
,
int
N
,
char
*
filename
);
int
main
()
{
srand
(
time
(
0
));
int
N
=
1
00
;
int
N
=
NB_G
*
2
00
;
vec
*
points
=
malloc
(
sizeof
(
vec
)
*
N
);
double
c
=
randNum
(),
d
=
randNum
();
...
...
@@ -23,26 +26,57 @@ int main() {
vec
tmp
;
tmp
.
x
=
randNum
();
r
=
randNum
()
/
10
0
.
0
;
r
=
randNum
()
/
10
.
0
;
tmp
.
y
=
c
*
tmp
.
x
+
d
+
r
;
points
[
i
]
=
tmp
;
}
vec
ab
;
ab
.
x
=
rand
()
%
10
;
ab
.
y
=
rand
()
%
10
;
int
*
rnd_indexes
=
rand_order_indexes
(
N
);
vec
**
groups
=
cut_table
(
rnd_indexes
,
N
,
points
);
vec
*
ab
=
malloc
(
sizeof
(
vec
)
*
NB_G
);
for
(
int
i
=
0
;
i
<
NB_G
;
i
++
)
{
ab
[
i
].
x
=
randNum
();
ab
[
i
].
y
=
randNum
();
int
other
=
i
+
1
,
to_test
=
i
-
1
;
if
(
other
>=
NB_G
)
{
other
=
0
;
}
else
if
(
to_test
<
0
)
{
to_test
=
NB_G
-
1
;
}
minimiser
(
groups
[
i
],
groups
[
other
],
N
/
NB_G
,
1e-6
,
&
ab
[
i
]);
printf
(
"model: a: %g, b: %g
\n
"
,
ab
[
i
].
x
,
ab
[
i
].
y
);
vec
test_ab
;
calc_ab
(
groups
[
to_test
],
NB_G
,
&
test_ab
);
printf
(
"test: a: %g, b: %g
\n\n
"
,
test_ab
.
x
,
test_ab
.
y
);
}
PrintVec
(
ab
);
//minimiser(points, N, 0.001, &ab);
printf
(
"c: %g, d: %g
\n
"
,
c
,
d
);
minimiser
(
points
,
N
,
0
.
001
,
&
ab
);
printf
(
"%g, %g
\n
"
,
c
,
d
);
for
(
int
i
=
0
;
i
<
NB_G
;
i
++
)
{
char
fname
[
11
];
sprintf
(
fname
,
"G%d.txt"
,
i
+
1
);
write_to_file
(
groups
[
i
],
N
/
NB_G
,
fname
);
}
write_to_file
(
ab
,
points
,
N
,
"points
.txt"
);
write_to_file
(
ab
,
NB_G
,
"ab
.txt"
);
system
(
"python3 graph.py"
);
free
(
ab
);
free
(
points
);
free
(
rnd_indexes
);
for
(
int
i
=
0
;
i
<
NB_G
;
i
++
)
{
free
(
groups
[
i
]);
}
free
(
groups
);
return
EXIT_SUCCESS
;
}
...
...
@@ -51,15 +85,43 @@ double randNum() {
return
rand
()
/
(
double
)
RAND_MAX
;
}
void
write_to_file
(
vec
ab
,
vec
*
points
,
int
N
,
char
*
filename
)
{
vec
**
cut_table
(
int
*
rand_indexes
,
int
N
,
vec
*
points
)
{
vec
**
groups
=
malloc
(
sizeof
(
vec
*
)
*
NB_G
);
int
frac
=
N
/
NB_G
;
for
(
int
i
=
0
;
i
<
NB_G
;
i
++
)
{
groups
[
i
]
=
malloc
(
sizeof
(
vec
)
*
frac
);
}
for
(
int
i
=
0
;
i
<
N
;
i
++
)
{
groups
[
i
/
frac
][
i
%
frac
]
=
points
[
rand_indexes
[
i
]];
}
return
groups
;
}
int
*
rand_order_indexes
(
int
N
)
{
int
*
indexes
=
calloc
(
0
,
sizeof
(
int
)
*
N
);
for
(
int
i
=
1
;
i
<
N
;
i
++
)
{
int
rnd_index
=
rand
()
%
N
;
while
(
indexes
[
rnd_index
]
!=
0
)
{
rnd_index
=
rand
()
%
N
;
}
indexes
[
rnd_index
]
=
i
;
}
return
indexes
;
}
void
write_to_file
(
vec
*
points
,
int
N
,
char
*
filename
)
{
FILE
*
fp
=
fopen
(
filename
,
"w"
);
if
(
fp
==
NULL
)
{
return
;
}
fprintf
(
fp
,
"%f,%f
\n
"
,
ab
.
x
,
ab
.
y
);
for
(
int
i
=
0
;
i
<
N
;
i
++
)
{
fprintf
(
fp
,
"%f,%f
\n
"
,
points
[
i
].
x
,
points
[
i
].
y
);
}
...
...
This diff is collapsed.
Click to expand it.
opti_num.c
+
27
−
10
View file @
80c0c057
...
...
@@ -3,14 +3,23 @@
#include
<math.h>
#include
<stdio.h>
void
minimiser
(
vec
*
vecs
,
int
N
,
double
lambda
,
vec
*
ab
)
{
void
minimiser
(
vec
*
vecs
1
,
vec
*
vecs2
,
int
N
,
double
lambda
,
vec
*
ab
)
{
double
sum_x
=
0
,
sum_xx
=
0
,
sum_xy
=
0
,
sum_y
=
0
;
for
(
int
i
=
0
;
i
<
N
;
i
++
)
{
sum_x
+=
vecs
[
i
].
x
;
sum_xx
+=
vecs
[
i
].
x
*
vecs
[
i
].
x
;
sum_xy
+=
vecs
[
i
].
x
*
vecs
[
i
].
y
;
sum_y
+=
vecs
[
i
].
y
;
for
(
int
i
=
0
;
i
<
2
*
N
;
i
++
)
{
if
(
i
<
N
)
{
sum_x
+=
vecs1
[
i
].
x
;
sum_xx
+=
vecs1
[
i
].
x
*
vecs1
[
i
].
x
;
sum_xy
+=
vecs1
[
i
].
x
*
vecs1
[
i
].
y
;
sum_y
+=
vecs1
[
i
].
y
;
}
else
{
int
tmp_i
=
i
%
N
;
sum_x
+=
vecs2
[
tmp_i
].
x
;
sum_xx
+=
vecs2
[
tmp_i
].
x
*
vecs2
[
tmp_i
].
x
;
sum_xy
+=
vecs2
[
tmp_i
].
x
*
vecs2
[
tmp_i
].
y
;
sum_y
+=
vecs2
[
tmp_i
].
y
;
}
}
vec
next_ab
=
*
ab
,
err_quad
;
...
...
@@ -19,15 +28,23 @@ void minimiser(vec *vecs, int N, double lambda, vec *ab) {
*
ab
=
next_ab
;
err_quad
.
x
=
ab
->
x
*
sum_xx
+
ab
->
y
*
sum_x
-
sum_xy
;
err_quad
.
y
=
ab
->
x
*
sum_x
+
N
*
ab
->
y
-
sum_y
;
err_quad
.
y
=
ab
->
x
*
sum_x
+
(
2
*
N
)
*
ab
->
y
-
sum_y
;
next_ab
=
Subtract
(
*
ab
,
Multiply
(
err_quad
,
lambda
));
printf
(
"%g, %g
\n
"
,
next_ab
.
x
,
next_ab
.
y
);
}
while
(
GetNorme
(
Subtract
(
next_ab
,
*
ab
))
>
EPSILON
);
}
void
calc_ab
(
double
sum_x
,
double
sum_xx
,
double
sum_xy
,
double
sum_y
,
int
N
,
vec
*
ab
)
{
void
calc_ab
(
vec
*
points
,
int
N
,
vec
*
ab
)
{
double
sum_x
=
0
,
sum_xx
=
0
,
sum_xy
=
0
,
sum_y
=
0
;
for
(
int
i
=
0
;
i
<
N
;
i
++
)
{
sum_x
+=
points
[
i
].
x
;
sum_xx
+=
points
[
i
].
x
*
points
[
i
].
x
;
sum_xy
+=
points
[
i
].
x
*
points
[
i
].
y
;
sum_y
+=
points
[
i
].
y
;
}
ab
->
x
=
(
sum_y
*
sum_x
-
N
*
sum_xy
)
/
(
sum_x
*
sum_x
-
N
*
sum_xx
);
ab
->
y
=
(
sum_y
-
ab
->
x
*
sum_x
)
/
N
;
...
...
This diff is collapsed.
Click to expand it.
opti_num.h
+
2
−
2
View file @
80c0c057
...
...
@@ -4,7 +4,7 @@
#define _OPTI_NUM_H_
#define EPSILON 1e-6
void
minimiser
(
vec
*
vecs
,
int
N
,
double
lambda
,
vec
*
ab
);
void
calc_ab
(
double
sum_x
,
double
sum_xx
,
double
sum_xy
,
double
sum_y
,
int
N
,
vec
*
ab
);
void
minimiser
(
vec
*
vecs
1
,
vec
*
vecs2
,
int
N
,
double
lambda
,
vec
*
ab
);
void
calc_ab
(
vec
*
points
,
int
N
,
vec
*
ab
);
#endif
This diff is collapsed.
Click to expand it.
Preview
0%
Loading
Try again
or
attach a new file
.
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Save comment
Cancel
Please
register
or
sign in
to comment