Scholarlink

The quality of iron ore is an essential factor in setting up an iron ore mine. The grade of an iron ore actually predicts it’s richness in iron. Accurate prediction of mineral grades is a fundamental step in mineral exploration and resource estimation.In this paper, a novel mechanism to predict an iron ore's richness has been proposed using Artificial Neural Network (ANN). A mathematical model consisting of three-layer feed-forward back propagation Multilayer perceptron to predict the iron ore’s richness has been applied and trained. A database consisting of 515 data sets recorded at the Taldih subdivision in the Sunder-Garh district of Orissa, India, is used to train and test the capability of the ANN model. Coefficient of determination (R2) and Mean Square Error (MSE) has been taken as the indicators of the network's performance. A satisfactory result has been found, and analysis has been done to justify the proposed work.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An Efficient Neural Network-Based Mathematical Modelling for Iron Ore Quality Prediction

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Anuradha Sharma,
  • Sudip Kumar Sahana

摘要

The quality of iron ore is an essential factor in setting up an iron ore mine. The grade of an iron ore actually predicts it’s richness in iron. Accurate prediction of mineral grades is a fundamental step in mineral exploration and resource estimation.In this paper, a novel mechanism to predict an iron ore's richness has been proposed using Artificial Neural Network (ANN). A mathematical model consisting of three-layer feed-forward back propagation Multilayer perceptron to predict the iron ore’s richness has been applied and trained. A database consisting of 515 data sets recorded at the Taldih subdivision in the Sunder-Garh district of Orissa, India, is used to train and test the capability of the ANN model. Coefficient of determination (R2) and Mean Square Error (MSE) has been taken as the indicators of the network's performance. A satisfactory result has been found, and analysis has been done to justify the proposed work.