chakon / parametric_modeling Goto Github PK
View Code? Open in Web Editor NEWMaster Thesis Work on iCub Humanoid Robot (Created Package for Parameteric Modeling)
Master Thesis Work on iCub Humanoid Robot (Created Package for Parameteric Modeling)
Same problem as in Issue #3. The selected lambda is not reliable.
kammo@kammo-Latitude-E5540:~/ICUB/Parametric_Modelling/Version0.1/build2$ ./iCubParis02_simple_analysis --dataset ../../Data_Sets/part1-left.csv --results results.csv --xtrain ../../Data_Sets/Xtr_2psmall.txt --ytrain ../../Data_Sets/ytr_2psmall.txt
1405
Specified Task Sequence :
Sequence of size:2
[Task 0: split]: ho... done.
[Task 1: paramsel]: hoprimal... done.
Save cycle...
[Task 0: split]: ho... saving
[Task 1: paramsel]: hoprimal... saving
Saving opt in iCubParis02_simple_analysis.bin
lambda values:
[ 0.00014291 0.00014291 112.15 0.00014291 0.00014291 0.00014291 ]
final lambda value: 0.00014291
================= Printing All ================
[ Name ] = iCubParis02_simple_analysis
[ calibfile ] = foo
[ combineclasses ] = Pointer to the function <mean> whose signature is: T (_func)(T_, int)
[ eig_percentage ] = 5
[ epochs ] = 4
[ hoperf ] = macroavg
[ hoproportion ] = 0.2
[ name ] = iCubParis02_simple_analysis
[ nholdouts ] = 1
[ nlambda ] = 20
[ nsigma ] = 25
[ paramsel ] =
~~~~~~~ GurlsOptionList: paramsel
[ Name ] = paramsel
[ guesses ] =
[ 0.00014291 0.000376804 0.000993502 0.00261952 0.00690675 0.0182107 0.0480153 0.126599 0.333799 0.88011 2.32054 6.11846 16.1323 42.5351 112.15 295.701 779.66 2055.69 5420.14 14291 ]
```
[ lambdas ] =
```
[ 0.00014291 0.00014291 112.15 0.00014291 0.00014291 0.00014291 ]
```
[ lambdas_round ] =
```
[ 0.00014291 0.00014291 112.15 0.00014291 0.00014291 0.00014291 ]
```
[ perf ] =
```
[ 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.428571 0.357143 0.107143 0.0357143 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.996047 0.996047 0.996047 0.996047 0.996047 0.996047 0.996047 0.996047 0.996047 0.996047 0.996047 0.996047 0.996047 0.996047 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ]
```
[ perfeval ] = acc
[ ploteval ] = acc
[ plotstr ] = iCubParis02_simple_analysis
[ predbagmethod ] = vote
[ processes ] =
~~~~~~~ GurlsOptionList: processes
[ Name ] = processes
[ one ] = ( ComputeNsave, ComputeNsave )
```
```
[ randfeats ] =
```
~~~~~~~ GurlsOptionList: randfeats
[ D ] = 500
[ Name ] = randfeats
[ samplesize ] = 100
```
[ saveanalysis ] = 1
[ savefile ] = iCubParis02_simple_analysis.bin
[ savekernel ] = 1
[ seq ] = Sequence of size:2
[ singlelambda ] = Pointer to the function <median> whose signature is: T (*func)(T*, int)
[ smallnumber ] = 1e-08
[ split ] =
~~~~~~~ GurlsOptionList: split
[ Name ] = split
[ indices ] =
gMat2D: (1405 x 1) matrix of type m
[ lasts ] =
[ 1124 ]
```
```
[ subsize ] = 50
[ time ] =
```
~~~~~~~ GurlsOptionList: elapsedtime
[ Name ] = elapsedtime
[ one ] =
[ 0 0.001 ]
```
[ tmpdir ] = iCubParis02_simple_analysis.bin
[ todisk ] = 1
[ verbose ] = 1
[ version ] = 2.0
~~~~~~~ ================= Printed All ================
```
After issuing the command:
./iCubParis02_simple_analysis --dataset ../../Data_Sets/part1-left.csv --results results.csv --xtrain ../../Data_Sets/Xtr_2psmall.txt --ytrain ../../Data_Sets/ytr_2psmall.txt
The program was executed and the results stored in results.csv.
However, the content of results.csv may not be saved properly. This is the file content, which includes several NaNs. How should it be interpreted?
-nan,-nan,-nan,-nan,-nan,-nan,14.3552,-3.7024,-2.86709,0.223453,1.02526,-0.205156,-nan,-nan,-nan,-nan,-nan,-nan,14.3652,-3.66704,-2.86271,0.221296,1.02564,-0.203341,-nan,-nan,-nan,-nan,-nan,-nan,14.3861,-3.58752,-2.85829,0.216157,1.02557,-0.199272,-nan,-nan,-nan,-nan,-nan,-nan,14.4177,-3.48059,-2.83147,0.209381,1.02497,-0.193797,-nan,-nan,-nan,-nan,-nan,-nan,14.4694,-3.2802,-2.80688,0.19641,1.02344,-0.18354,-nan,-nan,-nan,-nan,-nan,-nan,14.4893,-3.21236,-2.78233,0.192061,1.02225,-0.180068,-nan,-nan,-nan,-nan,-nan,-nan,14.5535,-2.96845,-2.71763,0.176628,1.01936,-0.167566,-nan,-nan,-nan,-nan,-nan,-nan,14.5767,-2.87325,-2.69536,0.170671,1.01847,-0.162681,-nan,-nan,-nan,-nan,-nan,-nan,14.6392,-2.62884,-2.60436,0.155183,1.01292,-0.150147,-nan,-nan,-nan,-nan,-nan,-nan,14.705,-2.329,-2.51795,0.136538,1.00776,-0.134744,-nan,-nan,-nan,-nan,-nan,-nan,14.7384,-2.16127,-2.47174,0.126229,1.00503,-0.12612,-nan,-nan,-nan,-nan,-nan,-nan,14.7947,-1.83051,-2.40164,0.106172,1.00125,-0.109095,-nan,-nan,-nan,-nan,-nan,-nan,14.8497,-1.46336,-2.31202,0.0841438,0.996355,-0.0901846,-nan,-nan,-nan,-nan,-nan,-nan,14.8935,-1.11957,-2.22089,0.0636656,0.990688,-0.072467,-nan,-nan,-nan,-nan,-nan,-nan,14.9275,-0.795893,-2.13034,0.0444654,0.983272,-0.0557769,-nan,-nan,-nan,-nan,-nan,-nan,14.6074,-0.605329,-
[FILE IS TOO LONG --> TRUNCATED]
-- Found YCM: /home/kammo/Repos/codyco-superbuild/build/install/share/YCM (found version "0.2.0~20141017+gita2285614")
Looking for KDL in: /home/kammo/Repos/codyco-superbuild/build/install
-- Using iCub from install
-- Configuring done
-- Generating done
-- Build files have been written to: /home/kammo/ICUB/Parametric_Modelling/Version0.1/build2
[ 11%] Building CXX object CMakeFiles/Data_Model_trail.dir/Data_Model_trail.cpp.o
/home/kammo/Repos/Parametric_Modeling/Version0.1/Data_Model_trail.cpp:1:0: warning: "NDEBUG" redefined [enabled by default]
#define NDEBUG
^
:0:0: note: this is the location of the previous definition
Linking CXX executable Data_Model_trail
/usr/bin/ld: cannot find -lkdl_codyco
/usr/bin/ld: cannot find -lkdl_format_io
/usr/bin/ld: cannot find -liDynTree
collect2: error: ld returned 1 exit status
make[2]: *** [Data_Model_trail] Error 1
make[1]: *** [CMakeFiles/Data_Model_trail.dir/all] Error 2
make: *** [all] Error 2
*** Failure: Exit code 2 ***
hoperf is set to macroavg in test_for_lambda, which is not suitable for regression problems.
See program output:
kammo@kammo-Latitude-E5540:~/ICUB/Parametric_Modelling/Version0.1/build2$ ./test_for_lambda ../../Data_Sets/Xtr_2psmall.txt ../../Data_Sets/ytr_2psmall.txt
=========== Number of rows for training ============
rows of training :1405
============Number of rows displaye d ==============
Specified Task Sequence :
Sequence of size:2
[Task 0: split]: ho... done.
[Task 1: paramsel]: loocvprimal... done.
Save cycle...
[Task 0: split]: ho... saving
[Task 1: paramsel]: loocvprimal... saving
Saving opt in test_for_lambda.bin
lambda values:
[ 0.000143623 0.000143623 297.176 0.000143623 0.000143623 0.000143623 ]
final lambda value:0.000143623
================= Printing All ================
[ Name ] = test_for_lambda
[ calibfile ] = foo
[ combineclasses ] = Pointer to the function <mean> whose signature is: T (_func)(T_, int)
[ eig_percentage ] = 5
[ epochs ] = 4
_[ hoperf ] = macroavg_
[ hoproportion ] = 0.2
[ name ] = test_for_lambda
[ nholdouts ] = 1
[ nlambda ] = 20
[ nsigma ] = 25
[ paramsel ] =
~~~~~~~ GurlsOptionList: paramsel
[ Name ] = paramsel
[ guesses ] =
[ 0.000143623 0.000378683 0.000998456 0.00263258 0.0069412 0.0183015 0.0482547 0.127231 0.335463 0.884499 2.33212 6.14897 16.2127 42.7472 112.71 297.176 783.548 2065.94 5447.17 14362.3 ]
```
[ lambdas ] =
```
[ 0.000143623 0.000143623 297.176 0.000143623 0.000143623 0.000143623 ]
```
[ perf ] =
```
[ 0.504274 0 0.992236 0 0 0
0.504274 0 0.992236 0 0 0
0.504274 0 0.992236 0 0 0
0.504274 0 0.992236 0 0 0
0.504274 0 0.992236 0 0 0
0.504274 0 0.992236 0 0 0
0.504274 0 0.992236 0 0 0
0.504274 0 0.992236 0 0 0
0.504274 0 0.992236 0 0 0
0.504274 0 0.992236 0 0 0
0.504274 0 0.992236 0 0 0
0.495726 0 0.993012 0 0 0
0.487179 0 0.994565 0 0 0
0.410256 0 0.995342 0 0 0
0.34188 0 0.999224 0 0 0
0.128205 0 1 0 0 0
0.00854701 0 1 0 0 0
0 0 1 0 0 0
0 0 1 0 0 0
0 0 1 0 0 0 ]
```
[ perfeval ] = acc
[ ploteval ] = acc
[ plotstr ] = test_for_lambda
[ predbagmethod ] = vote
[ processes ] =
~~~~~~~ GurlsOptionList: processes
[ Name ] = processes
[ one ] = ( ComputeNsave, ComputeNsave )
```
```
[ randfeats ] =
```
~~~~~~~ GurlsOptionList: randfeats
[ D ] = 500
[ Name ] = randfeats
[ samplesize ] = 100
```
[ saveanalysis ] = 1
[ savefile ] = test_for_lambda.bin
[ savekernel ] = 1
[ seq ] = Sequence of size:2
[ singlelambda ] = Pointer to the function <median> whose signature is: T (*func)(T*, int)
[ smallnumber ] = 1e-08
[ split ] =
~~~~~~~ GurlsOptionList: split
[ Name ] = split
[ indices ] =
gMat2D: (1405 x 1) matrix of type m
[ lasts ] =
[ 1124 ]
```
```
[ subsize ] = 50
[ time ] =
```
~~~~~~~ GurlsOptionList: elapsedtime
[ Name ] = elapsedtime
[ one ] =
[ 0 0.015 ]
```
[ tmpdir ] = test_for_lambda.bin
[ todisk ] = 1
[ verbose ] = 1
[ version ] = 2.0
~~~~~~~ ================= Printed All ================
```
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.