<div class="sectionTitle">Information Science Discussion Papers Series: 2009 Abstracts</div>

<hr>

<a name="#dp2009-01"></a><h3>2009/01: Tag based model for knowledge sharing in agent society</h3>
<h4>S. Savarimuthu, M. Purvis and M. Purvis</h4>

<p>In this paper we discuss a tag-based model that facilitates knowledge sharing in the context of agents playing the knowledge sharing game. Sharing the knowledge incurs a cost for the sharing agent, and thus non-sharing is the preferred option for selfish agents. Through agent-based simulations we show that knowledge sharing is possible even in the presence of non-sharing agents in the population. We also show that the performance of an agent society can be better when some agents bear the cost of sharing instead of the whole group sharing the cost.</p>

<p><strong>Keywords: </strong>cooperation,
altruism,
tags,
knowledge sharing,
multi-agent based simulation,
artificial society</p>

<p><a href="papers/dp2009-01.pdf">Download</a> (PDF, 424 KB)</p>

<hr> 

<a name="#dp2009-02"></a><h3>2009/02: Monitoring social expectations in Second Life</h3> 

<h4>S. Cranefield and G. Li</h4> 

<p>Online virtual worlds such as Second Life provide a rich medium for unstructured human interaction in a shared simulated 3D environment. However, many human interactions take place in a structured social context where participants play particular roles and are subject to expectations governing their behaviour, and current virtual worlds do not provide any support for this type of interaction. There is therefore an opportunity to adapt the tools developed in the MAS community for structured social interactions between software agents (inspired by human society) and adapt these for use with the computer-mediated human communication provided by virtual worlds.</p>

<p>This paper describes the application of one such tool for use with Second Life. A model checker for online monitoring of social expectations defined in temporal logic has been integrated with Second Life, allowing users to be notified when their expectations of others have been fulfilled or violated. Avatar actions in the virtual world are detected by a script, encoded as propositions and sent to the model checker, along with the social expectation rules to be monitored. Notifications of expectation fulfilment and violation are returned to the script to be displayed to the user. This utility of this tool is reliant on the ability of the Linden scripting language (LSL) to detect events of significance in the application domain, and a discussion is presented on how a range of monitored structured social scenarios could be realised despite the limitations of LSL.</p> 


<p><a href="papers/dp200902.pdf">Download</a> (PDF, 188 KB)</p> 


<hr>

<a name="#dp2009-03"></a><h3>2009/03: A software framework for application development using ZigBee protocol</h3>
<h4>B.T.R. Savarimuthu, M. Bruce and M. Purvis</h4>

<p>The problem with the uptake of new technologies such as ZigBee is the lack of development environments that help in faster application software development. This paper describes a software framework for application development using ZigBee wireless protocol. The architecture is based on defining XML based design interfaces that represent the profiles of ZigBee nodes that are used in the application.</p>


<p><a href="papers/dp2009-03.pdf">Download</a> (PDF, 168 KB)</p>

<hr> 

<a name="#dp2009-04"></a><h3>2009/04: Automatic sapstain detection in processed timber through image feature analysis</h3> 

<h4>J. D. Deng</h4> 

<p>Sapstain is considered a defect that must be removed from processed wood. So far, research in automatic wood inspection systems has been mostly limited to dealing with knots. In this paper, we extract a number of colour and texture features from wood pictures. These features are then assessed using machine learning techniques via feature selection, visualization, and finally classification. Apart from average colour and colour opponents, texture features are also found to be useful in classifying sapstain. This implies a significant modification to the domain understanding that sapstain is mainly a discolourization effect. Preliminary results are presented, with satisfactory classification performance using only a few selected features. It is promising that a real world wood inspection system with the functionality of sapstain detection can be developed.</p> 


<p><a href="papers/dp200904.pdf">Download</a> (PDF, 884 KB)</p> 

<hr> 

<a name="#dp2009-05"></a><h3>2009/05: Software effort estimation: Harmonizing algorithms and domain knowledge in an integrated data mining approach</h3> 

<h4>J. D. Deng and M. K. Purvis</h4> 

<p>Software development effort estimation is important for quality management in the software development industry, yet its automation still remains a challenging issue. Applying machine learning algorithms alone often can not achieve satisfactory results. In this paper, we present an integrated data mining framework that incorporates domain knowledge into a series of data analysis and modeling processes, including visualization, feature selection, and model validation. An empirical study on the software effort estimation problem using a benchmark dataset shows the effectiveness of the proposed approach.</p> 

<p><strong>Keywords:</strong> software effort estimation, machine learning</p> 

<p><a href="papers/dp2009-05.pdf">Download</a> (PDF, 184 KB)</p>