LIKELIHOOD FILTER FOR CLUSTER DETECTION

The likelihood filter for cluster detection introduced by Postman et a l. is generalized by using standard procedures and models originally d eveloped in the theory of point processes. It is shown that the filter formulae of Postman et al, can be recovered in cases where background fields dominate the number counts. The performance of the generalized method is illustrated by using Monte Carlo simulations and by analyzi ng galaxy distributions extracted from the COSMOS galaxy catalogue. Th e generalized method has the advantage of being less biased at the exp ense of some higher computational effort.