Abstract

Abstract
Portugal stands out as a recognized country in the production of superior quality wine, the two main reasons being the edaphoclimatic conditions and the grapevine heritage. To maximize quality it is important that the various steps of the winemaking process be submitted to effective control and monitoring. Since the alcoholic fermentation is a crucial stage of the winemaking process, this paper describes a low cost prototype to perform the supervision and control of the alcoholic fermentation process in a winery. To demonstrate the viability of the practical application of this solution in real conditions, a prototype was installed in the winery of Escola Superior Agraria de Coimbra (ESAC), to control the fermentation temperature of white must in a medium size vat.

Abstract
Decision rules are one of the most expressive languages for machine learning. In this paper we present Adaptive Model Rules (AMRules), the first streaming rule learning algorithm for regression problems. In AMRules the antecedent of a rule is a conjunction of conditions on the attribute values, and the consequent is a linear combination of attribute values. Each rule in AMRules uses a Page-Hinkley test to detect changes in the process generating data and react to changes by pruning the rule set. In the experimental section we report the results of AMRules on benchmark regression problems, and compare the performance of our algorithm with other streaming regression algorithms.

Abstract
Informed driving is becoming a key feature to increase the sustainability of taxi companies. Some recent works are exploring the data broadcasted by each vehicle to provide live information for decision making. In this paper, we propose a method to employ a learning model based on historical GPS data in a real-time environment. Our goal is to predict the spatiotemporal distribution of the Taxi-Passenger demand in a short time horizon. We did so by using learning concepts originally proposed to a well-known online algorithm: the perceptron [1]. The results were promising: we accomplished a satisfactory performance to output the next prediction using a short amount of resources.

Abstract

Abstract