Level of Vulnerability of Educational Institutions in Face El Nino Phenomenon and its Classification with the Neural Network

Abstract

The El Nino Phenomenon is a climatic event whose consequences are devastating for Peru (Landslides, floods, etc). Due to this, in the present investigation a neural network is proposed, this RNA has the capacity to evaluate the level of vulnerability before this event of a building, more specifically, an educational institution. An artificial multi-layer perceptron neural network was developed, trained with the backpropagation algorithm. This training was carried out using the results of the risk level assessment developed by the Ministry of Agriculture and Irrigation of Peru, approximately twelve thousand records were used. He developed two types of architectures with different number of neurons in the hidden layer. Finally, the first architecture was selected as the most suitable because it had a root error of 0.05, being less than the second architecture. With the training obtained in the first neuron, a web application was implemented to classify the level of vulnerability of an educational institution according to certain patterns present in it.

Keywords

• artificial neural nets
• vulnerability classifier
• el nino
• peru
• multilayer perceptron