دسته‌بندي ناپيوستگي‌هاي موجود در توده سنگ با استفاده از شبكه عصبي رقابتي در چهارچوب احتماالت‌

DISCONTINUITY CLUSTERING IN ROCK MASSES WITH THE APPLICATION OF KOHONEN NEURAL NETWORK UNDER THE FRAMEWORK OF PROBABILITIES

Distinct clustering of discontinuities is very useful for analyses and the design of the found a tion of st r uct ures or rock masses. A traditional method is conto uring a number of significan t pole densities plotted on the ste reonet . The use of different techniques for conto ur ing a poles plot generates different results. Therefore, in this paper , art ificial in telligen ce methods, with the a pplicatio n of probabilities , are used to cluster the discontinuities. One of th e advan tages of t his metho d is to determine the acc urate number of discontinuity cluster s, t he belonging of each discontinuity to a specific cluster and its depend ence on the distribu tion of dat a.