Skip to main content

Table 6 Wheat datasets

From: Genome-enabled prediction using probabilistic neural network classifiers

  MLP15% PNN15% MLP30% PNN30%
Upper class
GY-1 0.204 (0.084) 0.288 (0.140) 0.406 (0.113) 0.475 (0.102)
GY-2 0.270 (0.108) 0.307 (0.111) 0.485 (0.113) 0.567 (0.116)
GY-3 0.227 (0.114) 0.268 (0.108) 0.366 (0.100) 0.453 (0.118)
GY-4 0.242 (0.110) 0.325 (0.118) 0.409 (0.107) 0.518 (0.115)
GY-5 0.284 (0.115) 0.326 (0.142) 0.505 (0.116) 0.550 (0.107)
GY-6 0.504 (0.172) 0.561 (0.157) 0.637 (0.115) 0.701 (0.083)
GY-7 0.199 (0.091) 0.290 (0.117) 0.423 (0.114) 0.529 (0.115)
Lower class
DTH-1 0.304 (0.113) 0.414 (0.124) 0.522 (0.107) 0.630 (0.091)
DTH-2 0.297 (0.117) 0.429 (0.132) 0.433 (0.110) 0.521 (0.104)
DTH-3 0.364 (0.149) 0.511 (0.151) 0.547 (0.115) 0.650 (0.095)
DTH-4 0.254 (0.077) 0.298 (0.089) 0.297 (0.070) 0.363 (0.097)
DTH-5 0.275 (0.131) 0.384 (0.164) 0.440 (0.104) 0.546 (0.087)
DTH-6 0.380 (0.091) 0.467 (0.094) 0.465 (0.099) 0.520 (0.112)
DTH-7 0.368 (0.114) 0.482 (0.113) 0.521 (0.124) 0.591 (0.115)
DTH-8 0.264 (0.097) 0.382 (0.103) 0.452 (0.102) 0.599 (0.095)
DTH-9 0.261 (0.103) 0.367 (0.112) 0.416 (0.099) 0.535 (0.107)
DTH-10 0.447 (0.109) 0.553 (0.114) 0.462 (0.112) 0.578 (0.124)
  1. Mean values of the area under the precision-recall curve AUCpr (standard deviation in parentheses) of 50 random partitions for the 15 and 30 % upper classes for grain yield (GY) in 7 environments (1–7) and 15 and 30 % lower classes for days to heading (DTH) in 10 environments (1–10) for classifiers MLP and PNN. Numbers in bold are the highest AUCpr values between MLP and PNN for 15 and 30 %