Analysis of a Complex Dataset Using the Combined MST and Auto-Contractive Map
The auto-contractive map neural network is applied to the London Metropolitan Police drug trafficking database to identify previously unknown relationships and depicts the new knowledge in the form of a minimal spanning tree. Each subset of data is shown along with the corresponding minimal spanning tree, and a detailed description is presented to aid the reader in correctly interpreting the structure. The knowledge attained using the auto-contractive map can assist police agencies in directing limited resources to geographical areas and individuals that have the greatest probability of success. Profiles are created based entirely on the neural network algorithms that are independent of any human subjectivity.
KeywordsMinimum Span Tree Power Centre Territorial Adjacency Territorial Structure Global Graph
- Buscema, M., & Sacco, P. L. (2010). Auto-contractive maps, the H function, and the maximally regular graph (MRG): a new methodology for data mining (chapter 11). In V. Capecchi et al. (Eds.), Applications of mathematics in models, artificial neural networks and arts. Dordrecht/London: Springer. doi: 10.1007/978-90-481-8581-8_11.Google Scholar
- Buscema, M. (2007). Contractive Maps. Software for programming Auto Contractive Maps. (Semeion Software #15, v. 2), Rome.Google Scholar
- Buscema, M. (2007). Constraints Satisfaction Networks. Software for programming Non Linear Auto-Associative Networks (Semeion Software #14, v. 10), Rome.Google Scholar
- Buscema, M. (2008). MST. Software for programming Trees from artificial networks weights matrix (Semeion Software #38, v 5), Rome.Google Scholar
- Massini, G. (2007). Tree Visualizer. Software to draw and manipulate tree graph (Semeion Software #40, v. 3), Rome.Google Scholar