Consistency-oriented benchmarking or comparing agricultural enterprises

Abstract

The management of agricultural farms means the manipulation of ecological and economical systems, which inevitably needs the safe and effective integration of the knowledge that was accumulated on a social level (cf. online extension systems) before, to individual decision-making, because the ”polyhistor-like” approach is getting less viable. The essence of the consultancy process is an expertise containing the individual, ad hoc opinion of the expert, which is based on the expert’s personal experiences, and the attributes of the given farm, and thus it is an insecure judgment. One of the more sophisticated methods is an (online) expert system (like a chess automaton) , which selects one the ”standard phrases” of the experts, based on the farm data and the whole farm statistics (c.f. FADN). These standard phrases (templates) are selected using the verification methods that were developed by the experts themselves, and those are selected which fits the most to the given farm. For example, the results that are created when individual farms are compared (c.f. benchmarking) to the best farms may be inconsistent. The solution of this needs the integration of consistency checks in the consultancy process. The goal of the lecture and the article is the introduction of a new approach of this topic, and a short demonstration of the first results. The analytic method that is being refined, based on results of partial similarity analyses, seek the answer for the following questions: Does a given farm have any ecological comparative advantages and/or excellent management capabilities compared to other farms (with the similar size, production line, income-generating capability, and within the same region)? How much does the provided indicators contribute to the succes of the farms, both generally and farm-specific? How does the variation of the indicators affect the succes of the farming? What is the real maximum of the indicators that describe the succes of the farming? Which indicators are more important than the others in the different regional, farm-type based comparisons on the average? This new approach have two components: first is the objectivity-driven learning method, second is the explotitation of the dialectics of the parts and the holistics in order to minimalize the risk of the consultancy,and to enhance the consistency. The presented methodology is suitable to handle both static and dynamic problems. The demonstrated examples are based on the research database of the German Institute of Informatics.