Wrong error measure used in test_for_lambda

[raffaello-camoriano]

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

# New task sequence...

[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 ================
```