Expertise Analysis Support. The system developed by McDonald [8] in-
corporates a user-customisable expertise ¯ltering process. Such a process takes
a set of un¯ltered names recommended by the system, and then reorders and
removes items to generate a re¯ned recommendation. Support for this ¯ltering
process can be provided by accessing the database that maintains personal and
organisational relevant data, e.g. departmental a±liation.
A user seeking individuals as sources of information and/or as individuals
who can perform given organisational functions, imposes their own requirements
on the Expert Finder's functionality. Hence it is the users who should select the
appropriate ¯lters depending on their needs, and the system should only support
the expert ¯nding process by providing analysis functionality. This means that
including the ability to rank experts using di®erent user-customisable criteria
(rather than the mere provision of a linear listing based on pre-determined cri-
teria) can considerably enhance the expert ¯nding applications. For example,
a ¯lter which represents the working relationships among experts can be used
for catalysing collaboration in co-authoring documents or launching projects. In
cases where users require knowledge that can only be shared through the human
dimension, a ¯lter based on the physical distance between departments can be
employed.
Another approach that can support users in expertise analysis is to increase
the system's transparency. This can be achieved by providing interfaces to ac-
cess the expertise evidence, together with the expertise recognition logic. For
example, the system can supply the scores for ranking the experts along with
the associated expertise evidence. It allows users to evaluate the validity of the
system's recognition logic. This in turn permits the incorporation of functional-
ity, which can assist the users in evaluating and exploring the expertise evidence,
i.e. spotting anomalies in the expert ¯nding process, giving users greater trust
in the system