Prediction Model for Peruvian cocoa production using Deep Learning and Long Short-Term Memory techniques (#314)

Authors

Porras, Ronald

Ovalle, Christian

Abstract

Cocoa is an essential product of Peruvian agriculture since a large number of families depend on it. The International Cocoa Organization has suggested carrying out strategies in order to avoid overproduction and minimize the price crisis. In the research work, the development of a prediction model for the production of Peruvian cocoa was carried out using the deep learning and lstm techniques, since this will allow estimating production and avoiding the devaluation of the sale price. The PMBOK methodology was used, which has 5 phases, it will also allow control over the risks that may affect the project and greater efficiency in the quality of deliverables. The prediction model for cocoa production was made using a python programming language and the TensorFlow library, which makes it possible to implement models based on deep learning in an easier and faster way. It can be concluded that the prediction model helps to estimate the production of Peruvian cocoa using deep learning and lstm techniques with an accuracy percentage of 99.6%.