Occupational accident databases are widely used by Workers' Compensation Authorities and private Safety and prevention Management with different purposes. A systematic accidents reporting leads to a large and complex data base where each element is characterized by many parameters and dealing with this amount of information becomes hard. Data mining techniques represent an efficient toll for locate useful information from large databases and in the last 20 years several techniques have had a wide applications in many classification and analysis problems. Among these methods, a coupled clustering, constituted by a projection of data from high-dimensional space to a low dimensional space and a numerical clustering, was presented in 2009 and performed promising results. Carrying on with this approach, this paper introduce a new release of the method that allows to exceed some lacks regarding the numerical clustering stability and the result visualisation. The method has been applied successfully to a data base of occupational accident with the purpose of grouping the element according with their similarity and making a clear visualisation of this classification. The capability of this method of grouping data and visualize them represent a powerful toll for analyst that have to deal with large occupational data base and it can represent a flexible support in the preventive measurement designing.

Workplace accidents analysis with a coupled clustering methods: S.O.M. and K-means algorithms / Comberti, Lorenzo; Baldissone, Gabriele; Demichela, Micaela. - In: CHEMICAL ENGINEERING TRANSACTIONS. - ISSN 2283-9216. - STAMPA. - 43:(2015), pp. 1261-1266. (Intervento presentato al convegno ICHEAP12: 12TH INTERNATIONAL CONFERENCE ON CHEMICAL & PROCESS ENGINEERING    tenutosi a Milano, Italia nel MAY 19-22, 2015) [http://www.aidic.it/cet/15/43/211.pdf].

Workplace accidents analysis with a coupled clustering methods: S.O.M. and K-means algorithms

COMBERTI, LORENZO;BALDISSONE, GABRIELE;DEMICHELA, Micaela
2015

Abstract

Occupational accident databases are widely used by Workers' Compensation Authorities and private Safety and prevention Management with different purposes. A systematic accidents reporting leads to a large and complex data base where each element is characterized by many parameters and dealing with this amount of information becomes hard. Data mining techniques represent an efficient toll for locate useful information from large databases and in the last 20 years several techniques have had a wide applications in many classification and analysis problems. Among these methods, a coupled clustering, constituted by a projection of data from high-dimensional space to a low dimensional space and a numerical clustering, was presented in 2009 and performed promising results. Carrying on with this approach, this paper introduce a new release of the method that allows to exceed some lacks regarding the numerical clustering stability and the result visualisation. The method has been applied successfully to a data base of occupational accident with the purpose of grouping the element according with their similarity and making a clear visualisation of this classification. The capability of this method of grouping data and visualize them represent a powerful toll for analyst that have to deal with large occupational data base and it can represent a flexible support in the preventive measurement designing.
File in questo prodotto:
Non ci sono file associati a questo prodotto.
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11583/2651336
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo