Assessing Inhomogeneous Indicator-Related Typologies Through the Reverse. . .
43
original category of large cities, while the tenth – new – cluster is again a singleton,
but this time an extremely rich provincial commune, where large-scale mining and
industrial development is located. The original clusters are, however, even less well
reconstructed.
7 Final Observations
The approach here applied allowed for an easy reconstruction of the main axis of
partitioning, even if not all of the clusters obtained lie (exactly) along this axis.
It is easy to point out these of the original categories of municipalities, which –
confirming the quite obvious intuitive supposition – are not in conformity with this
axis. This is, largely, what we intended to obtain, when trying to check the capacity
of the approach to handle similar typologies, in which some of the types may not
necessarily be situated roughly along the “leading indicator”, whether explicit or
implicit. At the same time, we got the confirmation as to the primary hypothesis of
existence of the distinct subgroups, even if in a sense distributed along the “leading
indicator” axis. This motivates us to apply the approach to other, similar cases, in
which the interpretations, in terms of relation to the “main axis” or the “leading
indicator” may not be as simple as in this particular one.
On the substantive side only few conclusions, which appear to be truly justified
against the background of the results obtained, namely: (1) it is necessary to
consider separately Warsaw and its zone of influence; (2) more generally, treatment
of the suburban zones / zones of influence of bigger and smaller agglomerations
ought to be reconsidered in the direction of appropriate adjustment / identification
of the thresholds / criteria accounted for (a similar observation applies, anyway,
also to the rural farming communities); (3) regarding the “special types” of rural
municipalities, they should be better justified in terms of their more general socioeconomic characteristics.
References
Ester, M., Kriegel, H.-P., Sander, J., & Xu, X.-w. (1996). A density-based algorithm for discovering
clusters in large spatial databases with noise. In E. Simondis, J. Han, & U. M. Fayyad (Eds.),
Proceedings of the Second International Conference on Knowledge Discovery and Data Mining
(KDD-96) (pp. 226–231). Menlo Park: AAAI Press.
Hubert, L., & Arabie, P. (1985). Comparing partitions. Journal of Classification, 2(1), 193–218.
https://doi.org/10.1007/BF01908075.
Lance, G. N., & Williams, W. T. (1966). A generalized sorting strategy for computer classifications.
Nature, 212, 218.
Lance, G. N., & Williams, W. T. (1967). A general theory of classification sorting strategies. 1.
Hierarchical systems. Computer Journal, 9, 373–380.
Lloyd, S. P. (1957). Least squares quantization in PCM. Bell Telephone Labs Memorandum,
Murray Hill, NJ; reprinted in IEEE Trans. Information Theory, IT-28 (1982), 2, 129–137.
43
original category of large cities, while the tenth – new – cluster is again a singleton,
but this time an extremely rich provincial commune, where large-scale mining and
industrial development is located. The original clusters are, however, even less well
reconstructed.
7 Final Observations
The approach here applied allowed for an easy reconstruction of the main axis of
partitioning, even if not all of the clusters obtained lie (exactly) along this axis.
It is easy to point out these of the original categories of municipalities, which –
confirming the quite obvious intuitive supposition – are not in conformity with this
axis. This is, largely, what we intended to obtain, when trying to check the capacity
of the approach to handle similar typologies, in which some of the types may not
necessarily be situated roughly along the “leading indicator”, whether explicit or
implicit. At the same time, we got the confirmation as to the primary hypothesis of
existence of the distinct subgroups, even if in a sense distributed along the “leading
indicator” axis. This motivates us to apply the approach to other, similar cases, in
which the interpretations, in terms of relation to the “main axis” or the “leading
indicator” may not be as simple as in this particular one.
On the substantive side only few conclusions, which appear to be truly justified
against the background of the results obtained, namely: (1) it is necessary to
consider separately Warsaw and its zone of influence; (2) more generally, treatment
of the suburban zones / zones of influence of bigger and smaller agglomerations
ought to be reconsidered in the direction of appropriate adjustment / identification
of the thresholds / criteria accounted for (a similar observation applies, anyway,
also to the rural farming communities); (3) regarding the “special types” of rural
municipalities, they should be better justified in terms of their more general socioeconomic characteristics.
References
Ester, M., Kriegel, H.-P., Sander, J., & Xu, X.-w. (1996). A density-based algorithm for discovering
clusters in large spatial databases with noise. In E. Simondis, J. Han, & U. M. Fayyad (Eds.),
Proceedings of the Second International Conference on Knowledge Discovery and Data Mining
(KDD-96) (pp. 226–231). Menlo Park: AAAI Press.
Hubert, L., & Arabie, P. (1985). Comparing partitions. Journal of Classification, 2(1), 193–218.
https://doi.org/10.1007/BF01908075.
Lance, G. N., & Williams, W. T. (1966). A generalized sorting strategy for computer classifications.
Nature, 212, 218.
Lance, G. N., & Williams, W. T. (1967). A general theory of classification sorting strategies. 1.
Hierarchical systems. Computer Journal, 9, 373–380.
Lloyd, S. P. (1957). Least squares quantization in PCM. Bell Telephone Labs Memorandum,
Murray Hill, NJ; reprinted in IEEE Trans. Information Theory, IT-28 (1982), 2, 129–137.