Premium
The Application and Potential of Cluster Analysis in the Interpretation of multivariate particulate data
Author(s) -
Cooke Steven P.,
Butters Gordon
Publication year - 1984
Publication title -
particle and particle systems characterization
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.877
H-Index - 56
eISSN - 1521-4117
pISSN - 0934-0866
DOI - 10.1002/ppsc.19840010113
Subject(s) - cluster (spacecraft) , multivariate statistics , diagram , interpretation (philosophy) , scatter plot , multivariate analysis , field (mathematics) , range (aeronautics) , data mining , particle (ecology) , computer science , position (finance) , biological system , statistics , mathematics , materials science , geology , oceanography , pure mathematics , composite material , biology , programming language , finance , economics
The classification of powders into groups possessing specific behavioural identities has long been recognized as desirable. This has been approached by describing different powders using two variables and plotting the data on a scatter diagram. The property of interest is then inferred from a powder's position on the scatter diagram. Unfortunately, a powder's behaviour is rarely adequately described by only two variables. To describe a powder more fully requires more variables and this presents a problem in interpretation. Cluster analysis refers to a range of techniques for interpreting multivariate data sets and is used in many fields. This paper describes cluster analysis and considers the method and potential of its application to particle technology. By way of an illustration, a cluster analysis has been performed on twenty‐seven powders resulting in groups recognizable by their characteristic fluidisation and flow properties. It is concluded that cluster analysis could be a powerful tool in predictive and analytical work in the field of particle technology.