A Decision Support System for Vine Growers Based on a Bayesian Network

We propose here a decision support system for vine growers to assess the quality of a vineyard to be planted. The quality of a vineyard is defined by the probability of possible profitability of the wine sales he is able to produce. The model, based on a Bayesian network (BN), takes into account environment and the parameters defining vineyard status with their associated interactions. BN are widely used for knowledge representation and reasoning under uncertainty in natural resource management. There is a rising interest in BN as tools for ecological and agronomic modelling. Data were collected from knowledge of vine-growing experts. We developed a C# computer program predicting the likely quality of a vineyard. The model has been validated on existing vineyards with prediction ability around 75%. This system should ease assessments of the likely impact of the choices and decisions of vine growers on the quality of new vineyards to be planted in any part of the world. No such model has been developed before for vine growers. Supplementary materials accompanying this paper appear on-line.