Chapter Contents | Introduction | Authors | References
Table of Contents | Index | Previous Section | Next Section
After cluster analysis processing, the remaining small extended sources were tested for repeatability on the weeks/months timescales of weeks to months. Only those detections found on at least two hours-confirming coverages were accepted as reliable sources. Two sources were regarded as being weeks-confirmed if,
(a) they were in the same wavelength band
(b) they had detection times which differed by at least 36 hours;
(c) they satisfied a position confirmation criterion detailed below; and
(d) the ratio of the flux of the brighter detection to that of the fainter was less than 3.
The link parameter for weeks-confirmation was taken to be the (X)2 statistic of the two-dimensional position-match confidence of the two detections; in computing (X)2, each detection was modeled as a two-dimensional Gaussian distribution characterized by the extent matrix. Two detections positionally confirmed if their link parameter was less than 1.4 (see Section V.E.7.b).
When a group of two or more weeks-confirming sources was identified, it was merged into a single weeks-confirmed source. The flux was taken to be the average of the fluxes of the two sources. The centroid and extent matrix of the weeks-confirmed source were calculated and the extent matrix was diagonalized to yield semi-major and semi-minor axes and the twist angle between the semi-major axis and the ecliptic meridian. The number of hours-confirmed sightings that make up the weeks-confirmed source is called NCOVR in Section X.C.
If all the hours-confirmed sources contributing to a weeks-confirmed source positionally confirmed each other, then the source was said to be mutually confirmed. If an hours-confirmed source (denoted by A) weeks-confirmed with two or more other sources (denoted by Bl, B2, etc.) which did not mutually confirm, the following was done to extract the best weeks-confirmed source. First, a decision function, D, was evaluated for each of sources B 1, B2, etc., where
(V.E.6)
and 
2 is
the link parameter between sources A and B and R is
the ratio of the brighter flux to that of the fainter.
The source for which D was the smallest was found and the ratio D/Dmin was found for each of the sources Bl, B2, etc. All sources for which D/Dmin was above a threshold of 2 were discarded. If only one source remained then this was chosen as the weeks-confirming partner of source A. Otherwise, the source was considered confused and the candidates were rejected.
Chapter Contents | Introduction | Authors | References
Table of Contents | Index | Previous Section | Next Section