Article: Simulation of copper ore formation based on multi-source geological data and deep learning Journal: International Journal of Information and Communication Technology (IJICT) 2026 Vol.27 No.90 pp.71 - 92 Abstract: Accurate prediction of copper ore mineralisation potential remains a challenging task in mineral resource exploration. To address the limitations of existing methods in characterising complex mineralisation processes and their unsatisfactory predictive performance, this paper proposes a novel deep learning framework with three key innovations. First, multi-source geological data are pre-processed using an entropy-weighted principal component analysis strategy to remove redundant features. Second, a meanshift-improved synthetic minority over-sampling technique is applied to oversample the training set, effectively balancing positive and negative sample ratios. Third, asymmetric convolution modules and a multi-spectral channel attention module are integrated into the ResNeXt50 network, significantly enhancing the model's ability to extract features from small-sized, multi-channel mineralisation data. Experimental results demonstrate that the proposed method achieves a prediction accuracy of 95.14%, outperforming the best benchmark method by at least 3.03%, and enables efficient simulation of copper ore mineralisation. Inderscience Publishers - linking academia, business and industry through research

Title: Simulation of copper ore formation based on multi-source geological data and deep learning

Authors: Jiulin Fan; Teng Wang

Addresses: Central South Institute of Metallurgical Geology, Yichang, 443000, China ' Central South Institute of Metallurgical Geology, Yichang, 443000, China

Abstract: Accurate prediction of copper ore mineralisation potential remains a challenging task in mineral resource exploration. To address the limitations of existing methods in characterising complex mineralisation processes and their unsatisfactory predictive performance, this paper proposes a novel deep learning framework with three key innovations. First, multi-source geological data are pre-processed using an entropy-weighted principal component analysis strategy to remove redundant features. Second, a meanshift-improved synthetic minority over-sampling technique is applied to oversample the training set, effectively balancing positive and negative sample ratios. Third, asymmetric convolution modules and a multi-spectral channel attention module are integrated into the ResNeXt50 network, significantly enhancing the model's ability to extract features from small-sized, multi-channel mineralisation data. Experimental results demonstrate that the proposed method achieves a prediction accuracy of 95.14%, outperforming the best benchmark method by at least 3.03%, and enables efficient simulation of copper ore mineralisation.

Keywords: multi-source geological data; deep learning; synthetic minority over-sampling technique.

DOI: 10.1504/IJICT.2026.155735

International Journal of Information and Communication Technology, 2026 Vol.27 No.90, pp.71 - 92

Received: 19 Apr 2026
Accepted: 01 Jun 2026
Published online: 11 Aug 2026 *

Title: Simulation of copper ore formation based on multi-source geological data and deep learning

Keep up-to-date