Optimisation of the filtration process by applicationof data mining

摘要

Beer filtration is a process of outstanding relevance still showing unrevealed influences. Tornoptimise filtration, novel data mining methods are used to analyse filtration and detect thosernunrevealed influences, which are often hidden in data usually collected in breweries. In thisrnapproach, filtration are classified by cluster methods, which form the basis for examiningrnlaboratory data and their influences. First results indicate an important impact of storage timernon filterability and prove the applicability of data mining methods for solving problemsrnduring the brewing process. The developed first models allow a prediction of overallrntendencies of filterability with a certainty up to 80 %.