Ensemble Random Weights Neural Network based Online Prediction Model of the Production Rate for Mineral Beneficiation Process

Abstract

Mineral beneficiation process consists of a series of unit processes, and it is of great significance to build the relationship between the technical indexes of each unit and the global production index. In this paper, we propose a novel online prediction model of the concentrate ore production rate in mineral beneficiation process based on the ensemble neural network with random weights. The model uses random vector functional link (RVFL) networks as the base components and employs the regularized negative correlation learning (RNCL) strategy to train the ensemble neural network with a collaborative approach, which takes the diversity and complexity of the model into account. For the selection of prediction model inputs, we introduce the maximal information coefficient (MIC) to analyze the relationships between the production rate and the inputs, i.e. the technical indexes of each unit. The variables which have larger relevance with the production rate are chosen as the inputs of the model. The experiments using the practical data of mineral beneficiation process are conducted. The experimental results demonstrate that the proposed prediction model outperforms the existing methods.