融合全监督学习的半监督矿石粒度预测算法
姜志宏,陈澳

Semi-supervised Ore Granularity Prediction Algorithm Incorporating Fully Supervised Learning
Zhihong JIANG,Ao CHEN
表4 高置信度伪标签样本数据(部分数据)
Table 4 High-confidence pseudo-label sample data(partial data)

矿石粒级

/mm

图像识别矿石粒度分布/%加权算术平均粒度/mm标准差

偏差

系数

模型预测矿石粒度分布/%
+5~100.332.5661.1610.1700.318
-50.3570.8220.3960.0600.377
+100.3663.4681.5040.2210.285
-50.2960.8670.3830.0550.393
+100.3113.1671.6060.2540.297
-50.3890.7920.4010.0630.279
+100.3053.1561.6000.2540.302
+5~100.3052.3671.2000.1900.416
+5~100.3212.5151.1760.1750.350
-50.3320.8380.3920.0590.318
+100.3663.4681.5300.2210.386
+5~100.3402.6011.1480.1660.280
-50.3600.8030.3880.0600.324
+100.3103.3631.5070.2240.402
+5~100.3802.5221.1300.1680.282
-50.3100.8410.3770.0560.305
+100.3103.3131.5350.2320.415
-50.2500.9030.3670.0580.325
+5~100.3802.6061.1140.1600.396
+5~100.3902.5691.1140.1630.325
+100.3003.1501.5960.2530.388
-50.3900.7880.3990.0630.282