Document Type : Original Research Paper
Authors
1 Department of Geospatial Information Systems, Faculty of Geodesy and Geomatics, K.N.Toosi University of Technology, Tehran, Iran
2 Department of Computer Science, Faculty of Mathematics and Computer Science, Shahid Chamran University of Ahvaz, Ahvaz, Iran
Abstract
Background and Objectives: Geospatial Information Systems provide required science and tools for various transactions related to spatial data. Such an ability caused use of Geospatial Information Systems in several fields. One of these fields is mine exploration which can be known as highly dependent field to the geology. On the other hand, the traditional methods are not suitable for requirements of mine exploration. Thus, the new methods are utilized more and more for mining as well as mine exploration. Artificial intelligence is one of the new sciences transforming human life. Artificial Neural Networks as a long-standing method of artificial intelligence have found many applications in mine exploration. The goal of this research is producing mineral potential map using artificial neural networks.
Methods: Artificial neural network is a classifier method in essence. This method acts like a black box which is trained at first, then it is capable of classifying the new data. In this research the data captured from exploration studies of porphyry copper deposit located in Yazd, Iran is used for producing mineral potential map. The mineral potential map is essential and very important for mining activities. The data is entered to a feedforward back propagation artificial neural network. The artificial neural network is used in two manners. The first manner is the usual one: The artificial neural network is trained by the aid of boreholes data and then it is used to forecast the potential of copper at every cell of the study area. In the second manner after training the network, its inner weights are extracted. These weights show the amount of importance that the neurons of the network have been considered for the input criteria. These weights are entered into the index overlay method. Afterwards, the criteria maps are combined by index overlay method and the mineral potential map is produced.
Findings: In this research more than testing the power of artificial neural networks in producing mineral potential map, testing the accuracy of its inner weights to be used in another method is aimed. At the training phase the values of criteria maps (essentially produced by exploration studies) at the positions of boreholes are entered into the artificial neural network and the network should forecast the potential of copper at that position qualitatively, while the true values are known from the data of boreholes. Then, the trained network forecast the potential of copper at every position of the deposit. The results revealed that the accuracy of artificial neural network when ignoring one of the non-efficient criteria can be reached up to 100 percent. However, the accuracy of index overlay method using the weights extracted from the artificial neural network is about 70 percent at maximum.
Conclusion: The results of this study revealed again the power of artificial neural networks in classification and combination of spatial data. Despite, the unique result of this research is that the inner weights of artificial neural network have the maximum performance in their network and using them for weighting and combining data by another method would not be useful. However, these weights can illustrate the order of importance of data.
Keywords
- Mine Exploration
- Copper Deposit
- Geospatial Information System
- Artificial Neural Network
- Data Combination
Main Subjects
COPYRIGHTS
© 2023 The Author(s). This is an open-access article distributed under the terms and conditions of the Creative Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) (https://creativecommons.org/licenses/by-nc/4.0/)
)