Free Access

Table 2.

Statistics of photometric redshift performance obtained for KiDS DR3 experiments with ANNz2 and MLPQNA vs. BPZ.

Sample Method Mean of δz = zphzsp Mean of δz/(1 + zsp) St.dev. of δz/(1 + zsp) SMADa of δz/(1 + zsp) % of outliers |δz|/(1 + zsp) > 0.15
Random subsample ANNz2 −3.3 × 10−3 3.3 × 10−3 0.073 0.026 3.5%
z〉 = 0.332 MLPQNA −2.0 × 10−3 3.9 × 10−3 0.079 0.026 3.4%
BPZb −1.9 × 10−2 −1.5 × 10−3 0.089 0.035 4.1%
Random 10% of r < 20, ANNz2 −2.4 × 10−3 8.3 × 10−3 0.102 0.034 7.1%
all from r ≤ 20, MLPQNA −3.2 × 10−3 7.2 × 10−3 0.116 0.034 7.4%
z〉 = 0.489 BPZb −5.8 × 10−2 −1.9 × 10−2 0.120 0.042 8.4%
Trained w/o COSMOS, ANNz2 −4.4 × 10−2 1.2 × 10−2 0.183 0.091 25.0%
tested on COSMOS ANNz2wc −6.7 × 10−2 −4.6 × 10−4 0.184 0.086 22.7%
z〉 = 0.784 MLPQNA −8.0 × 10−2 −2.7 × 10−3 0.204 0.086 23.6%
BPZb −2.4 × 10−1 −8.5 × 10−2 0.195 0.085 24.5%
Trained w/o CDFS, ANNz2 3.0 × 10−2 5.2 × 10−2 0.232 0.108 25.7%
tested on CDFS ANNz2wc 3.9 × 10−2 5.4 × 10−2 0.206 0.101 26.0%
z〉 = 0.742 MLPQNA 1.0 × 10−2 3.8 × 10−2 0.222 0.100 25.8%
BPZb −1.9 × 10−2 −7.2 × 10−2 0.183 0.083 23.7%

Notes. Results for the particular tests are provided in blocks of rows. See text for details.

(a)

SMAD is the scaled median absolute deviation, converging to standard deviation for Gaussian distributions.

(b)

BPZ is independent of the training sets – the numbers are given for comparison (for the same test samples). These statistics are based on the KiDS pipeline solution.

(c)

Training data weighted with the kNN method, weights propagated throughout the training and evaluation procedure.

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.