A novel statistical approach to assess the quality and commercial viability of a retail branded perishable fruit

Few studies have investigated sensory quality in multi-varietal fruit brands due to challenges in making comparisons between varieties harvested sequentially throughout the season. Sensory data is collected in crisp form and analysed using a great amount of numerical data. However, human perception is vague and an assessor's opinion comes in linguistic form. The purpose was to assess the sensory quality of branded apricot using a novel statistical approach. Physicochemical and sensory attributes were determined for two consecutive commercial seasons. A Linear mixed model, Multiple Factor Analysis (MFA) and fuzzy model were applied in order to assess quality stability and to create a prediction model. The brand was affected by the presence of different cultivars and the MFA demonstrated aroma and juiciness as the most differentiated parameters, while univariate analysis highlighted hardness and mealiness. Use of expert knowledge mixed with information from sample data enabled the development of a fuzzy model. Further investigation with a consumer test is needed in order to confirm results.