A cluster-graph model for herd characterisation in dairy farms equipped with an automatic milking system. (March 2018)

• Record Type: Journal Article
• Authors: Bonora, Filippo; Benni, Stefano; Barbaresi, Alberto; Tassinari, Patrizia; Torreggiani, Daniele
• Is Part Of: Biosystems engineering. Volume 167(2018)
• Page Start: 1
• Page End: 7
• Publication Date: 2018-03
• Subjects: Precision livestock farming -- Productivity -- Dairy cow -- Cluster; AMS automatic milking system -- ICT information and communications technology -- PLF precision livestock farming; Bioengineering -- Periodicals; Agricultural engineering -- Periodicals; Biological systems -- Periodicals; Génie rural -- Périodiques; Systèmes biologiques -- Périodiques; 631
• DOI: 10.1016/j.biosystemseng.2017.12.007
• Languages: English
• ISSNs: 1537-5110

Abstract

Abstract : The analysis of data recorded by Automatic Milking System (AMS) in dairy livestock barns has a great potential for herd management and farm building design. A big amount of data about milk production and cow welfare is available from milking robot and many researches are focussing on them in order to find relationships and correlations among the various parameters. The goal of the study is to develop and test an innovative procedure for the comprehensive analysis of AMS-generated multi-variable time-series, with a focus on herd segmentation, aiming to support dairy livestock farm management. In particular, the study purpose is to develop and test a cluster-graph model using AMS-generated data, designed to provide an automatic grouping of the cows based on production and behavioural features. First, a k -means cluster analysis has been implemented to the average of the time series of the main parameters recorded for each cow by AMS in a barn in Italy over a summer period. Then, all the resulting subgroups have been converted in a network and a cluster-graph analysis has been applied in order to find herd-descriptive subgraphs. The results of the study have the potential impact of improving herd characterisation and lending support to cow monitoring and management. Furthermore, this method could represent a feasible procedure to convert alphanumeric data in a simple graphic visualisation of the herd without losing the quantitative information about every single animal. Graphical abstract: Highlights: A cluster-graph method has been developed to analyse datasets from AMS. The method identified groups of animals with different conditions. Clusters of cows with clearly different average productivity were obtained. The method is fit to support cow monitoring and herd management.