An Overview of KASBAH

My current research is investigating the link between husbandry regime and otter behaviour for Asian Small-Clawed Otters (Aonyx cinereus). It forms part of a larger framework, called the KASBAH project, looking at how this relationship might be manipulated in order to produce desirable behaviour patterns and reduce undesirable ones (such as stereotypies), especially for unfamiliar exotics. It is hoped that the end result will be a tool that the less experienced keeper or institution can use to optimally tune their regimes, but that is a long way down the line. I am an honorary research associate at Bristol Vet School, working with Professor John Webster and Dr Mike Mendl.

The actual mechanics of a visit are as follows:

I like to arrive as soon as keepers get on site, and stay until they leave. I stand by the otter enclosure, and make focal observations at one minute intervals throughout the day. When the otters have their afternoon snooze (hoping that they do!), I have about an hour's chat with their primary caregiver to discuss their husbandry regime, and the reasoning behind it. If anyone else is interested in the otters, I would like to speak to them too!  These observations and recorded facts are then stored electronically and become the base data for KASBAH.  I visit each set of otters at least three times at different seasons (and hence weather, visitor numbers etc).  The whole visit follows a formal, written protocol, so that like can be compared with like.

The KASBAH Framework

The essence of the KASBAH proposal is the construction of a knowledge-based system (KBS) that can capture a wide range of quantitative and qualitative information about the Asian Short-Clawed Otter (Amblonyx cinereus). The information sources will include wild and captive behavioural studies, captive animal welfare information (some of which may be case study, informal observational or anecdotal in nature) and physiological results.

Such a KBS involves the integration of information (often by deriving a set of underlying rules or by identifying patterns in the data) for presentation to a range of users. Typically the corpus of information is expected to be dynamic and growing rather than a static entity best handled by a conventional data-base approach. However, a pragmatic approach will be adopted: this is not a specific technology looking for an application. User interaction with the system is expected, for convenience, to be via a Web front-end for both input and output.

The overall approach proposed has been successfully used at the Laboratory for a range of application areas- to date addressing physical or engineering rather than behavioural systems. The approach is iterative: moving towards a refinement of the system in conjunction with users. Following an initial specification and system architectural design stage, an initial corpus will be collected. This corpus will be derived from such sources as books, journals broadcast and other media, interviews (including informal conversation and anecdotal evidence). The interviews will be with experienced observers of the otter in the wild, successful keepers, animal behaviour / welfare specialists.

A variety of IT tools (including pattern recognition, data-mining, case-based reasoning and classification techniques) will be used to derive "patterns" or "rules" in the corpus that can be built up as an empirical model of "the otter". This model can be used, with its Web-based front-end, as a general retrieval system ("what is known about X") as well as a more specific environment assessment system (derived from work at the Laboratory on explanation based systems).

The main aims of the proposed KASBAH work are:

• to provide a computing system that will systematise a large and varied corpus of (primarily behavioural) information about the otter into a useful tool to provide appropriate information and knowledge for a range of end users (including zoologists, veterinarians and professional keepers);
• to satisfy an expressed need from professional animal keepers for such a tool: with the primary objective being improvement of the conditions of captive otters;
• to use a range of computing methodologies to produce an exemplar information system for biological applications, with the longer-term aim of producing a methodology with an associated technology that can be used for other species and application areas.

If this process is successful for Otters, it will be repeated for a dissimilar species, the Zebra, to give some indication as to whether it will be more generally useful.