Wyniki wyszukiwania - BazTech

Ograniczanie wyników

Znaleziono wyników: 2

Liczba wyników na stronie

Wyniki wyszukiwania

Wyszukiwano:
w słowach kluczowych: validity index

Sortuj według:

Ogranicz wyniki do:

Delineation of potential seismic sources using weighted K means cluster analysis and particle swarm optimization (PSO)

Sheikhhosseini Zohreh, Mirzaei Noorbakhsh, Heidari Reza, Monkaresi Hamed

Acta Geophysica

2021

Vol. 69, no. 6

2161--2172

Potential seismic sources play an important role in seismic hazard analysis. Identification of seismic sources is generally carried out on the basis of expert judgments, and in most cases, different and controversial results are obtained when several experts are consulted. In fact, the method of source identification is probably an important cause of uncertainty in the seismic hazard analysis. The main objective of this research is to provide an algorithm which combines the weighted K-means clustering analysis and Particle Swarm Optimization in order to automatically identify global optimum clusters by analysing seismic event data. These clusters, together with seismotectonic information, can be used to determine seismic sources. Two validity indexes, Davies–Bouldin's measure and Chou–Su–Lai's measure (CS), are used to determine optimum number of clusters. Study area is located at the longitude of 46°–48° E and latitude of 34°–36° N that is considered as the most seismically active part of Zagros continental collision zone, which has experienced large and destructive earthquakes due to movements of Sahneh and Nahavand segments of Zagros Main Recent Fault. As a result, 7-cluster model which is identified on the basis of DB validity index seems to be suitable for the considered earthquake catalogue, despite some limitations in partitioning.

Eksploracja medycznych regułowych baz wiedzy

Nowak-Brzezińska A., Rybotycki T.

Studia Informatica

2016

Vol. 37, nr 1

149--162

Celem pracy jest eksploracja (grupowanie i wizualizacja) medycznych regułowych baz wiedzy. W artykule opisano narzędzie CluVis, zaimplementowane przez autorów, pozwalające analizować (grupować przy użyciu hierarchicznej analizy skupień) reguły i wizualizować (przy użyciu tzw. map prostokątów) ich skupienia. W ramach eksperymentów przeanalizowano wpływ miar podobieństwa wewnątrz- i międzygrupowego, metod wizualizacji, a także miary jakości skupień na wyniki eksploracji (wykrycie tendencji, nietypowości w danych).

In this work the topic of applying clustering as a knowledge extraction method from real-world medical data is discussed. The authors propose hierarchical clustering method and visualization techniques for knowledge base representation in the context of medical knowledge bases for which data mining methods are success-fully employed and may resolve different problems. What is more, the authors analyze the impact of different clustering parameters on the result of searching through such structure.