Abstract
An attempt is made to study mathematical properties of singular value decomposition (SVD) and its data exploring capacity and to apply them to make exploratory type clustering for 10 climatic variables and thirty weather stations in Bangladesh using a newly developed graphical technique. Findings in SVD and Robust singular value decomposition (RSVD) based graphs are compared with that of classical K-means cluster analysis, its robust version, partition by medoids (PAM) and classical factor analysis, and the comparison clearly demonstrates the advantage of SVD over its competitors. Lastly the method is tested on well known Hawkins-Bradu-Kass (1984) data.
Highlights
The conventional approach to the singular value decomposition (SVD) requires that the matrix X be complete
In this article we mainly develop a graphical technique on the basis of SVD and Robust Singular Value Decomposition (RSVD), and apply to a data set of Bangladesh that contain 10 climatic variables of 30 principal weather stations with a view to clustering them, and compare its results with that of classical K–means cluster analysis, partitioning around medoid (PAM) method and classical factor analysis
We see that our exploratory technique serves the purpose of both the techniques- K-means cluster analysis, and classical factor analysis have shown that this same graph with a slight change could be effectively used for outliers detection both for supervised and unsupervised learning
Summary
The source of this data is the report entitled “Land Resources Appraisal of Bangladesh for Agricultural Development”, BGD/81/035, Technical Report 3, volume I, @ FAO 1988. This report was prepared for the Government of the People’s Republic of Bangladesh, based on the work of H. Antoine (Data Base Management Expert) and A. T. van Velthuizen (Land resources and Agricultural Consultants) of UNDP (United Nations Development Programme). This data set contains the values of 10 climatic variables for 30 principal stations. For 30 principal stations, data sets of average values of annually means of the following parameters were collated from the unpublished and published records of the Bangladesh Meteorological Department (BMD).
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.