Integrated methods of artificial intelligence in identifying promising gold mineralization zones in Zailik, northwest of Iran

In this article, artificial intelligence methods and geological evidence were used to identify promising areas for gold mineralization. Gold grade values in the Zailik prospecting area located in the northwest of Iran were estimated by two artificial intelligence methods, artificial neural network and its integration with particle swarm optimization algorithm, and the type of constituent rocks and alterations of the studied area was chosen as geology parameters. After the summary of expert opinions in geosciences and mining, the geological parameters were weighed, and to score the artificial intelligence methods for estimating gold geochemical values, the coefficient of determination and the root mean square error function were used. All these methods were compared to the analytical hierarchy process (AHP) in expert choice software. The highest score among geological parameters, after summarizing expert opinions related to lithology and also among artificial intelligence methods, due to the higher coefficient of determination and lower error function, is the combined method of artificial neural network with particle swarm optimization algorithm was awarded. Finally, in the Arc Gis software, all the mentioned methods are united by the fuzzy overlay method. According to the final presented modeling, the northern and northeastern parts of the reconnaissance area are considered areas prone to gold mineralization to continue exploration was suggested.