Diego Hueltes, a Spanish computer engineer from Alcalá la Real (Jaén), has developed three models for predicting the price of olive oil using artificial intelligence.
"There are automatic learning algorithms that learn about past price fluctuations which, together with meteorological and production data, is capable of making an estimate close to reality," said Mr Hueltes, a computer engineer specialist in Big Data and Machine Learning, a branch of artificial intelligence that serves, among other things, to create prediction models.
Hueltes applied the latest techniques of artificial intelligence research to olive oil as "it occurred to me that olive oil is a good topic, related to my land and I hope in the future it can be used and its benefits have repercussions in the province".
The computer engineer, based in Marbella, explained how there are external factors that affect and cannot be predicted, but when the price depends solely on production, these algorithms are able to see the hidden relationships between the different variables to generate their estimates.
In this research, Hueltes has created several models and one of the "most interesting" is a model of "Deep learning" that simulates the behaviour of neurons and the human brain to learn to predict, in this case, the price of extra virgin olive oil.
He used data from the Andalusian Council on the price of olive oil in Jaén, the State Meteorological Agency (AEMET) and production of the Ministry of Agriculture, Fisheries and Food (MAPA).
Specifically, it has developed three models: one that predicts the price in the following week, another that predicts the price in four weeks and another that simply says whether the price will go up or down.
As detailed, there is a 2.9% absolute mean error for the prediction at four weeks; of 0.9% error for the prediction of the following week; and 76% accuracy in predicting the price direction. That 76% of success has been tested with the period between July 2017 and July 2018, with a 40% annual simulated benefit. More