XGBoost and LSTM predict next-shift hauling payload with median absolute errors around 14 to 15 percent, improving to 8.4 percent when scheduled-truck counts are added, though the unique contribution of simulated breakdown data is not demonstrated.
Designed to remember long-term patterns, LSTMs are ideal for mining fleet data, where cyclic production schedules are common
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Deep learning for predicting hauling fleet production capacity under uncertainties in open pit mines using real and simulated data
XGBoost and LSTM predict next-shift hauling payload with median absolute errors around 14 to 15 percent, improving to 8.4 percent when scheduled-truck counts are added, though the unique contribution of simulated breakdown data is not demonstrated.