Projects at LAM

12/18/2019

05:55 PM CFHTLS_z Action #5498 (New): CFHTLS training tests (suite)

+*PDF peakiness :*+
!PDFSAMPLE_ilt235_TEST.png! !PDFSAMPLE_igt235_TEST.png!
!PEAKINESS_TEST.png! Marie Treyer

05:50 PM CFHTLS_z Action #5497 (New): CFHTLS training tests with high resolution photo-z added

+*TRAINING SAMPLE:*+
Stephane added high resolution photo-z + lower resolution spec-z to the initial SPEC only cat... Marie Treyer

10/23/2019

05:53 PM CFHTLS_z Action #5433 (New): HSC_CLAUDS results compilation

!ZSPEC_MAG_TRAINING.png! !ZSPEC_MAG_TEST.png!
!ZSPEC_ZCNN_TRAINING.png! !ZSPEC_ZCNN_TEST.png!
!PIT_TRAINING.png! !P... Marie Treyer

04:41 PM CFHTLS_z Action #5432 (New): CFHTLS results compilation

!ZSPEC_MAG_TRAINING.png! !ZSPEC_MAG_TEST.png!
!ZSPEC_ZCNN_TRAINING.png! !ZSPEC_ZCNN_TEST.png!
!PIT_TRAINING.png! !P... Marie Treyer

09/18/2019

02:22 PM CFHTLS_z Action #5414 (New): more optimization: PDF median instead of PDF mean

Unlike in the SDSS, here the PDF median does better than our usual PDF mean.
!Fig2.png! !Fig3.png! !Fig4.png!
... Marie Treyer

09/13/2019

03:23 PM CFHTLS_z Action #5402 (New): Correction to improve low zspec bias (misclassified low z, faint galaxies)

WidzPeak=PDFwidth/zPeak, where zPeak is the highest peak of the (smoothed) PDF
This parameter efficiently picks ou... Marie Treyer

09/10/2019

11:01 AM CFHTLS_z Action #5401: Results with "correct" weights to reproduce observed mag/z distribution

Comparison between first and last CNN for W1_GOODPHOT (2.6M galaxies w/o zspec)
!nz_W1.jpg!
!zcnn_w_nw.jpg! Marie Treyer

09/09/2019

04:35 PM CFHTLS_z Action #5401: Results with "correct" weights to reproduce observed mag/z distribution

*Comparison between first and last training (the training samples were not exactly the same, the first one contained ... Marie Treyer

09/05/2019

03:18 PM CFHTLS_z Action #5401 (New): Results with "correct" weights to reproduce observed mag/z distribution

!zdis_train.jpg!
!zdis_test.jpg!
!zdis_smooth.jpg! Marie Treyer

06/14/2019

05:36 PM CFHTLS_z Action #5329 (New): CNN TEST WITH MAGNITUDE AND Z WEIGHTS

The training set is underpopulated at z<0.2 and I>20 (Fig 2), among other things. In the test sample (Primus) on the... Marie Treyer