Lalilti, A. and El Moumni, B. and El Arrim, A. and El Malki, H. and Alaoui Mhammedi, N. and Ousmana, H. and Berrada, M. and El Hmaidi, A. (2017) Prediction of carbonates, biogenic and detritic elements by artificial neural networks in the deposits of the north moroccan atlantic margin mud volcanoes Prédiction des carbonates, des éléments biogènes et détritiques par les réseaux de neurones artificiels dans les dépôts des volcans de boue de la marge atlantique nord marocaine. Bulletin de l'Institut Scientifique, Section Sciences de la Terre (39). pp. 121-134.

Full text not available from this repository.
Official URL: https://www.scopus.com/inward/record.uri?eid=2-s2....

Abstract

This work presents a prediction attempt of carbonate and detrital and biogenic elements in depending on granulometric size and mineralogical parameters in the late Quaternary deposits of mud volcanoes in the Gulf of Cadiz and the North Moroccan Atlantic margin. It is a kind of modeling based on the method of Artificial Neural Networks "RNA" type PMC (multilayer perceptrons). It was carried out with the Matlab language that represents a non-recurrent multilayer network based on a supervised learning and the Levenberg Marquardt algorithm. The results, very good, show that the three parameters studied were predicted with high accuracy using a neural model of configuration 8-4-3 with a non-linear activation function sigmoid type for the layer hidden and a linear activation function of Purelin in the output layer. Neural networks, very performant, showed significant ability of learning and prediction of carbonates and Biogenic and clastic elements with a correlation coefficient of 0.98 to 0.99, and a very low mean square error from 0.035 to 0.037 for the database used in addition to a better choice of the architecture of the neural network realized through preliminary tests. This shows that the parameters studied in the data set are associated with the modeled parameters by a non-linear relationship. © 2017 Universite Mohammed V, Institut Scientifique.

Item Type: Article
Subjects: Earth and Planetary Sciences
Divisions: SCIENTIFIC PRODUCTION > Earth and Planetary Sciences
Depositing User: Administrateur Eprints Administrateur Eprints
Last Modified: 31 Jan 2020 15:46
URI: http://eprints.umi.ac.ma/id/eprint/2861

Actions (login required)

View Item View Item