Are statistics and machine learning enough to make predictions and forecasts?

Abstract

Currently the techniques used to predict the future are statistical and machine learning techniques. The first continues the trend of historical data. The second learns from previous cases training. Both use historical information but do not take in mind key factors that can make the final result change. A knowledge-based framework is presented that allows predictions of some kind ofevents to be made using artificial intelligence techniques. This requires an expert to enter the key factors that can change the trend of historical data into the system.The current framework has been applied prior to happening to two use cases, obtaining good preliminary results, in the framework of the developing of a PhD Thesis.