Importance The article addresses binarization and clusterization of arbitrary objects' state matrices on the case of export competitiveness indicators in the ‘Fresh food’ sector of sub-Saharan African countries. Objectives The purpose of the study is to develop a methodology of matrix clustering and to test it by solving the problem of spatial economic analysis. Methods The article presents a methodology for matrix clusterization of objects that involves creating objects' state matrix, multi-criteria threshold binarization of state matrix and clustering the resulting binary matrix into sub-matrices with different density of zero or single elements. Results The developed matrix clustering methodology was tested on export competitiveness indicators in the ‘Fresh food’ sector of sub-Saharan African countries. Conclusions The developed matrix clustering methodology tested on export competitiveness indicators in the ‘Fresh food’ sector of sub-Saharan African countries shows that most of these countries are found in the second cluster (the number of zero elements in binary matrix ranges from 25 to 50%) and the third cluster (the number of zero elements in binary matrix ranges from 50 to 75%). The competitiveness of the ‘Fresh food’ sector grew from the fourth cluster to the first one. It is shown that binary matrix clustering is rather resistant to changes in threshold criteria in the initial state matrix.
Oyanagi S., Kubota K., Nakase A. Matrix Clustering: A New Data Mining Method for CRM. Trans. IPSJ, 2001, vol. 42(8), pp. 2156–2166.
Oyanagi S., Kubota K., Nakase A. Application of Matrix Clustering to Web Log Analysis and Access Prediction. In: WEBKDD 2001 – Mining Web Log Data Across All Customers Touch Points, Third International Workshop, 2001, pp. 13–21.
Oyanagi S., Kubota K., Nakase A. Mining WWW Access Sequence by Matrix Clustering. In: WEBKDD 2002 – Mining Web Data for Discovering Usage Patterns and Profiles, LNAI 2703, 2003, pp. 119–136.
Sudhamathy G., Venkateswaran C.J. Matrix Based Fuzzy Clustering for Categorizatioin on Web Users and Web Pages. International Journal of Computer Applications, 2012, vol. 43, no. 14, pp. 43–47. URL: https://doi.org/10.5120/6175-8602
Kuo J.J., Zhang Y.J. A Library Recommender System Using Interest Change over Time and Matrix Clustering. Taipei, Taiwan, 2012, pp. 259–268.
Nagaraj G., Sheik Syed Abuthahir S., Manimaran A., Paramasamy S. Comparison of Matrix Clustering Methods to Design Cell Formation. International Journal of Applied Engineering Research, 2015, vol. 10, iss. 28, pp. 21900–21904.
Zhang Z.Y, Li T., Ding C., Ren X.W., Zhang S. Binary Matrix Factorization for Analyzing Gene Expression Data. Data Mining and Knowledge Discovery, 2010, vol. 20, iss. 1, pp. 28–52. URL: https://doi.org/10.1007/s10618-009-0145-2
Oyelade J., Isewon I., Oladipupo F. et al. Clustering Algorithms: Their Application to Gene Expression Data. Bioinformatics and Biology Insights, 2016, vol. 10, pp. 237–253.
Moskovkin V.M., Casimiro H. Klasterizatsiya mnogomernykh ob"ektov razlichnoi prirody: postanovka issledovatel'skoi zadachi. V kn.: Sovremennye problemy sotsial'no-ekonomicheskikh sistem v usloviyakh globalizatsii [Clustering the multidimensional objects of different nature: A statement of research problem. In: Modern problems of socio-economic systems under globalization]. Belgorod, NRU BelSU Publ., 2017, pp. 27–30.
Moskovkin V.M., Casimiro H. [Matrix clustering as a clustering of matrices of the same dimension]. Nauchnye vedomosti Belgorodskogo gosudarstvennogo universiteta. Ser.: Ekonomika. Informatika = Belgorod State University Scientific Bulletin. Economics. Information Technologies, 2017, no. 23, iss. 24, pp. 123–127.