Ali A. Alesheikh
Assistant Professor, Dept. of GIS Eng.
Faculty of Geodesy and Geomatics Eng., University of K.N.Toosi
Tehran, Iran
Email: [email protected]
Jalal Karami
PhD Student
University of K.N.Toosi University, Iran
Vahid Mazaheri
Consultant, Iran
Abstract:
A Geospatial Information System (GIS) is a decision support tool, since the results of all GIS activities are to provide information for managers to take optimum actions. The advances in computer sciences provided a suitable infrastructure to gather varieties of parameters in a model to analyze or predict scenarios. The significance of the parameters and their contributions in an environmental model is the subject of many researches. Many algorithms have been proposed to tackle the issues of decision making. They can be classified in two categories: knowledge driven and data driven. There are many data driven algorithms such as Weight of Evidence, Characteristic Analysis (CA) and Canonical Favorability Analysis (CFA). The weight and importance of data and evidences are extracted for combining and modeling the specified phenomena.
In this paper, methods for data combination and modeling are reviewed. The advantages and drawbacks of each method are evaluated. Artificial Neural Networks (ANN) is used to simulate the action of the human's neuron. The capabilities of ANN in comparison with other data driven algorithms in the field of mineral potential mapping have been evaluated. Mineral indices have been used to train the algorithm. In comparison to other algorithms, ANN showed a better results.
