Table 7 Large wheat data set 3
IBCFMF, K = 2MF, K = 3MF, K = 4
Env–TraitMeanPCLMeanPCLMeanPCLMeanPCL
GY_Year_14_150.33333.300.31133.300.37433.30.35333.3
GY_Year_15_16_1yb0.33533.800.25433.800.24133.80.24033.8
GY_Year_15_16_2yb0.30533.750.23433.750.19833.750.19733.75
GY_Year_16_17_1yb0.28528.30−0.17828.30−0.17828.30−0.17928.30
GY_Year_16_17_2yb0.28628.150.19328.150.22228.150.23028.15
GY_Year_16_17_3yb0.28728.250.14828.250.19328.250.19628.25
Average0.30530.930.16030.930.17530.930.1730.93
HD_Year_14_150.62753.650.48353.650.55653.650.56553.65
HD_Year_15_16_1yb0.50848.350.32048.350.61648.350.60748.35
HD_Year_15_16_2yb0.53749.450.44049.450.56749.450.56349.45
HD_Year_16_17_1yb0.65047.550.33547.550.56447.550.57147.55
HD_Year_16_17_2yb0.64647.700.55447.700.60247.70.61147.7
HD_Year_16_17_3yb0.63747.200.58947.200.65547.20.65647.2
Average0.60148.980.45448.980.5948.980.6048.98
DMT_Year_14_150.64749.500.42749.500.54649.500.56649.50
DMT_Year_15_16_1yb0.49139.550.51039.550.69739.550.69439.55
DMT_Year_15_16_2yb0.63648.850.46048.850.59648.850.62548.85
DMT_Year_16_17_1yb0.52737.300.27037.300.58237.300.58037.30
DMT_Year_16_17_2yb0.56640.350.49440.350.61140.350.61140.35
DMT_Year_16_17_3yb0.54838.900.56538.900.59238.900.58838.90
Average0.56942.410.45442.410.6042.410.6142.41
PH_Year_14_150.01520.300.08820.000.04420−0.01420
PH_Year_15_16_1yb0.02620.400.06920.40−0.01020.400.05720.4
PH_Year_15_16_2yb0.04421.350.09221.350.05121.350.05821.35
PH_Year_16_17_1yb0.21727.600.22627.600.22527.60.18727.6
PH_Year_16_17_2yb0.24329.250.22929.250.22829.250.19729.25
PH_Year_16_17_3yb0.23828.950.16628.950.24128.950.22328.95
Average0.13124.640.14524.590.1324.590.1224.59
Lodging_Year_14_150.34740.700.10440.700.12040.700.12440.70
Lodging_Year_15_16_1yb−0.19511.950.02431.450.06031.450.07431.45
Lodging_Year_15_16_2yb−0.2749.650.09937.250.11537.250.14037.25
Average−0.04120.770.07636.4670.09836.470.1136.47
  • Prediction accuracies with Pearson correlation for all genotypes missing in the years: GY_Year_14_15, GY_Year_15_16, and GY_Year_16_17 for the wheat data set under cross-validation scheme CV2. Here, only the IBCF and MF methods were implemented. Method MF was implemented with three values of latent features (K = 2, 3, 4). Traits in 1 yr were predicted with data from 1 yr before (1yb), 2 yr before (2yb), and 3 yr before (3yb). PCL denotes the percentage of common lines in the top 2000 lines. MF, matrix factorization.