Mining trading partners’ preferences for efficient multi-issue bargaining in E‑business

Classical negotiation models are weak in supporting real-world business negotiations because these models often assume that the preference information of each negotiator is made public. Although parametric learning methods have been proposed for acquiring the preference information of negotiation opponents, these methods suffer from the strong assumptions about the specific utility function and negotiation mechanism employed by the opponents. Consequently, it is difficult to apply these learning methods to the heterogeneous negotiation agents participating in e‑marketplaces. This paper illustrates the design, development, and evaluation of a nonparametric negotiation knowledge discovery method which is underpinned by the well-known Bayesian learning paradigm. According to our empirical testing, the novel knowledge discovery method can speed up the negotiation processes while maintaining negotiation effectiveness. To the best of our knowledge, this is the first nonparametric negotiation knowledge discovery method developed and evaluated in the context of multi-issue bargaining over e‑marketplaces.

Mining Negative Relevance Feedback for Information Filtering

It is a big challenge to clearly identify the boundary between positive and negative streams. Several attempts have used negative feedback to solve this challenge; however, there are two issues for using negative relevance feedback to improve the effectiveness of information filtering. The first one is how to select constructive negative samples in order to reduce the space of negative documents. The second issue is how to decide noisy extracted features that should be updated based on the selected negative samples. This paper proposes a pattern mining based approach to select some offenders from the negative documents, where an offender can be used to reduce the side effects of noisy features. It also classifies extracted features (i.e., terms) into three categories: positive specific terms, general terms, and negative specific terms. In this way, multiple revising strategies can be used to update extracted features. An iterative learning algorithm is also proposed to implement this approach on RCV1, and substantial experiments show that the proposed approach achieves encouraging performance.

An effective model of using negative relevance feedback for information filtering

Over the years, people have often held the hypothesis that negative feedback should be very useful for largely improving the performance of information filtering systems; however, we have not obtained very effective models to support this hypothesis. This paper, proposes an effective model that use negative relevance feedback based on a pattern mining approach to improve extracted features. This study focuses on two main issues of using negative relevance feedback: the selection of constructive negative examples to reduce the space of negative examples; and the revision of existing features based on the selected negative examples. The former selects some offender documents, where offender documents are negative documents that are most likely to be classified in the positive group. The later groups the extracted features into three groups: the positive specific category, general category and negative specific category to easily update the weight. An iterative algorithm is also proposed to implement this approach on RCV1 data collections, and substantial experiments show that the proposed approach achieves encouraging performance.

State-of-the-Art Review on Opinion Mining from Online Customers’ Feedback

Dealing with the ever-growing information overload in the Internet, Recommender Systems are widely used online to suggest potential customers item they may like or find useful. Collaborative Filtering is the most popular techniques for Recommender Systems which collects opinions from customers in the form of ratings on items, services or service providers. In addition to the customer rating about a service provider, there is also a good number of online customer feedback information available over the Internet as customer reviews, comments, newsgroups post, discussion forums or blogs which is collectively called user generated contents. This information can be used to generate the public reputation of the service providers’. To do this, data mining techniques, specially recently emerged opinion mining could be a useful tool. In this paper we present a state of the art review of Opinion Mining from online customer feedback. We critically evaluate the existing work and expose cutting edge area of interest in opinion mining. We also classify the approaches taken by different researchers into several categories and sub-categories. Each of those steps is analyzed with their strength and limitations in this paper.

Rough set-based reasoning and pattern mining for information filtering

An information filtering (IF) system monitors an incoming document stream to find the documents that match the information needs specified by the user profiles. To learn to use the user profiles effectively is one of the most challenging tasks when developing an IF system. With the document selection criteria better defined based on the users’ needs, filtering large streams of information can be more efficient and effective. To learn the user profiles, term-based approaches have been widely used in the IF community because of their simplicity and directness. Term-based approaches are relatively well established. However, these approaches have problems when dealing with polysemy and synonymy, which often lead to an information overload problem. Recently, pattern-based approaches (or Pattern Taxonomy Models (PTM) [160]) have been proposed for IF by the data mining community. These approaches are better at capturing sematic information and have shown encouraging results for improving the effectiveness of the IF system. On the other hand, pattern discovery from large data streams is not computationally efficient. Also, these approaches had to deal with low frequency pattern issues. The measures used by the data mining technique (for example, “support” and “confidences”) to learn the profile have turned out to be not suitable for filtering. They can lead to a mismatch problem. This thesis uses the rough set-based reasoning (term-based) and pattern mining approach as a unified framework for information filtering to overcome the aforementioned problems. This system consists of two stages - topic filtering and pattern mining stages. The topic filtering stage is intended to minimize information overloading by filtering out the most likely irrelevant information based on the user profiles. A novel user-profiles learning method and a theoretical model of the threshold setting have been developed by using rough set decision theory. The second stage (pattern mining) aims at solving the problem of the information mismatch. This stage is precision-oriented. A new document-ranking function has been derived by exploiting the patterns in the pattern taxonomy. The most likely relevant documents were assigned higher scores by the ranking function. Because there is a relatively small amount of documents left after the first stage, the computational cost is markedly reduced; at the same time, pattern discoveries yield more accurate results. The overall performance of the system was improved significantly. The new two-stage information filtering model has been evaluated by extensive experiments. Tests were based on the well-known IR bench-marking processes, using the latest version of the Reuters dataset, namely, the Reuters Corpus Volume 1 (RCV1). The performance of the new two-stage model was compared with both the term-based and data mining-based IF models. The results demonstrate that the proposed information filtering system outperforms significantly the other IF systems, such as the traditional Rocchio IF model, the state-of-the-art term-based models, including the BM25, Support Vector Machines (SVM), and Pattern Taxonomy Model (PTM).

Applying data mining to assess crash risk on curves

The wide range of contributing factors and circumstances surrounding crashes on road curves suggest that no single intervention can prevent these crashes. This paper presents a novel methodology, based on data mining techniques, to identify contributing factors and the relationship between them. It identifies contributing factors that influence the risk of a crash. Incident records, described using free text, from a large insurance company were analysed with rough set theory. Rough set theory was used to discover dependencies among data, and reasons using the vague, uncertain and imprecise information that characterised the insurance dataset. The results show that male drivers, who are between 50 and 59 years old, driving during evening peak hours are involved with a collision, had a lowest crash risk. Drivers between 25 and 29 years old, driving from around midnight to 6 am and in a new car has the highest risk. The analysis of the most significant contributing factors on curves suggests that drivers with driving experience of 25 to 42 years, who are driving a new vehicle have the highest crash cost risk, characterised by the vehicle running off the road and hitting a tree. This research complements existing statistically based tools approach to analyse road crashes. Our data mining approach is supported with proven theory and will allow road safety practitioners to effectively understand the dependencies between contributing factors and the crash type with the view to designing tailored countermeasures.

A role mining inspired approach to representing user behaviour in ERP systems

Despite all attempts to prevent fraud, it continues to be a major threat to industry and government. Traditionally, organizations have focused on fraud prevention rather than detection, to combat fraud. In this paper we present a role mining inspired approach to represent user behaviour in Enterprise Resource Planning (ERP) systems, primarily aimed at detecting opportunities to commit fraud or potentially suspicious activities. We have adapted an approach which uses set theory to create transaction profiles based on analysis of user activity records. Based on these transaction profiles, we propose a set of (1) anomaly types to detect potentially suspicious user behaviour and (2) scenarios to identify inadequate segregation of duties in an ERP environment. In addition, we present two algorithms to construct a directed acyclic graph to represent relationships between transaction profiles. Experiments were conducted using a real dataset obtained from a teaching environment and a demonstration dataset, both using SAP R/3, presently the most predominant ERP system. The results of this empirical research demonstrate the effectiveness of the proposed approach.

The nature and extent of policing alcohol related crime and reducing violence in and around late night entertainment areas

The misuse of alcohol is well documented in Australia and has been associated with disorders and harms that often require police attention. The extent of alcohol-related incidents requiring police attention has been recorded as substantial in some Australian cities (Arro, Crook, & Fenton, 1992; Davey & French, 1995; Ireland & Thommeny, 1993). A significant proportion of harmful drinking occurs in and around licensed premises (Jochelson, 1997; Stockwell, Masters, Phillips, Daly, Gahegan, Midford, & Philp, 1998; Borges, Cherpitel, & Rosovsky, 1998) and most of these incidents are not reported to police (Bryant & Williams, 2000; Lister, Hobbs, Hall, & Winlow, 2000). Alcohol-related incidents have also been found to be concentrated in certain places at certain times (Jochelson, 1997) and therefore manipulating the context in which these incidents occur may provide a means to prevent and reduce the harm associated with alcohol misuse. One of the major objectives of the present program of research was to investigate the occurrence and resource impact of alcohol-related incidents on operational (general duties) policing across a large geographical area. A second objective of the thesis was to examine the characteristics and temporal/spatial dynamics of police attended alcohol incidents in the context of Place Based theories of crime. It was envisaged that this approach would reveal the patterns of the most prevalent offences and demonstrate the relevance of Place Based theories of crime to understanding these patterns. In addition, the role of alcohol, time and place were also explored in order to examine the association between non criminal traffic offences and other types of criminal offences. A final objective of the thesis was to examine the impact of a situational crime prevention strategy that had been initiated to reduce the violence and disorder associated with late-night liquor trading premises. The program of research in this doctorate thesis has been undertaken through the presentation of published papers. The research was conducted in three stages which produced six manuscripts, five of which were submitted to peer reviewed journals and one that was published in a peer reviewed conference proceedings. Stage One included two studies (Studies 1 & 2) both of which involved a cross sectional approach to examine the prevalence and characteristics of alcohol-related incidents requiring police attendance across three large geographical areas that included metropolitan cities, provincial regions and rural areas. Stage Two of the program of research also comprised two cross sectional quantitative studies (Studies 3 & 4) that investigated the temporal and spatial dynamics of the major offence categories attended by operational police in a specific Police District (Gold Coast). Stage Three of the program of research involved two studies (Studies 5 & 6) that assessed the effectiveness of a situational crime prevention strategy. The studies employed a pre-post design to assess the impact on crime, disorder and violence by preventing patrons from entering late-night liquor trading premises between 3 a.m. and 5 a.m. (lockout policy). Although Study Five was solely quantitative in nature, Study Six included both quantitative and qualitative aspects. The approach adopted in Study Six, therefore facilitated not only a quantative comparison of the impact of the lockout policy on different policing areas, but also enabled the processes related to the implementation of the lockout policy to be examined. The thesis reports a program of research involving a common data collection method which then involved a series of studies being conducted to explore different aspects of the data. The data was collected from three sources. Firstly a pilot phase was undertaken to provide participants with training. Secondly a main study period was undertaken immediately following the pilot phase. The first and second sources of data were collected between 29th March 2004 and 2nd May 2004. Thirdly, additional data was collected between the 1st April 2005 and 31st May 2005. Participants in the current program of research were first response operational police officers who completed a modified activity log over a 9 week period (4 week pilot phase & 5 week survey study phase), identifying the type, prevalence and characteristics of alcohol-related incidents that were attended. During the study period police officers attended 31,090 alcohol-related incidents. Studies One and Two revealed that a substantial proportion of current police work involves attendance at alcohol-related incidents (i.e., 25% largely involving young males aged between 17 and 24 years). The most common incidents police attended were vehicle and/or traffic matters, disturbances and offences against property. The major category of offences most likely to involve alcohol included vehicle/traffic matters, disturbances and offences against the person (e.g., common & serious assaults). These events were most likely to occur in the late evenings and early hours of the morning on the weekends, and importantly, usually took longer for police to complete than non alcohol-related incidents. The findings in Studies Three and Four suggest that serious traffic offences, disturbances and offences against the person share similar characteristics and occur in concentrated places at similar times. In addition, it was found that time, place and incident type all have an influence on whether an incident attended by a police officer is alcohol-related. Alcohol-related incidents are more likely to occur in particular locations in the late evenings and early mornings on the weekends. In particular, there was a strong association between the occurrence of alcohol-related disturbances and alcohol-related serious traffic offences in regards to place and time. In general, stealing and property offences were not alcohol-related and occurred in daylight hours during weekdays. The results of Studies Five and Six were mixed. A number of alcohol-related offences requiring police attention were significantly reduced for some policing areas and for some types of offences following the implementation of the lockout policy. However, in some locations the lockout policy appeared to have a negative or minimal impact. Interviews with licensees revealed that although all were initially opposed to the lockout policy as they believed it would have a negative impact on business, most perceived some benefits from its introduction. Some of the benefits included, improved patron safety and the development of better business strategies to increase patron numbers. In conclusion, the overall findings of the six studies highlight the pervasive nature of alcohol across a range of criminal incidents, demonstrating the tremendous impact alcohol-related incidents have on police. The findings also demonstrate the importance of time and place in predicting the occurrence of alcohol-related offences. Although this program of research did not set out to test Place Based theories of crime, these theories were used to inform the interpretation of findings. The findings in the current research program provide evidence for the relevance of Place Based theories of crime to understanding the factors contributing to violence and disorder, and designing relevant crime prevention strategies. For instance, the results in Studies Five and Six provide supportive evidence that this novel lockout initiative can be beneficial for public safety by reducing some types of offences in particular areas in and around late-night liquor trading premises. Finally, intelligent-led policing initiatives based on problem oriented policing, such as the lockout policy examined in this thesis, have potential as a major crime prevention technique to reduce specific types of alcohol-related offences.

Application of text mining in analysing road crashes for road asset management

Traffic safety is a major concern world-wide. It is in both the sociological and economic interests of society that attempts should be made to identify the major and multiple contributory factors to those road crashes. This paper presents a text mining based method to better understand the contextual relationships inherent in road crashes. By examining and analyzing the crash report data in Queensland from year 2004 and year 2005, this paper identifies and reports the major and multiple contributory factors to those crashes. The outcome of this study will support road asset management in reducing road crashes.

Personalised ontology learning and mining for web information gathering

Over the last decade, the rapid growth and adoption of the World Wide Web has further exacerbated user needs for e±cient mechanisms for information and knowledge location, selection, and retrieval. How to gather useful and meaningful information from the Web becomes challenging to users. The capture of user information needs is key to delivering users' desired information, and user pro¯les can help to capture information needs. However, e®ectively acquiring user pro¯les is di±cult. It is argued that if user background knowledge can be speci¯ed by ontolo- gies, more accurate user pro¯les can be acquired and thus information needs can be captured e®ectively. Web users implicitly possess concept models that are obtained from their experience and education, and use the concept models in information gathering. Prior to this work, much research has attempted to use ontologies to specify user background knowledge and user concept models. However, these works have a drawback in that they cannot move beyond the subsumption of super - and sub-class structure to emphasising the speci¯c se- mantic relations in a single computational model. This has also been a challenge for years in the knowledge engineering community. Thus, using ontologies to represent user concept models and to acquire user pro¯les remains an unsolved problem in personalised Web information gathering and knowledge engineering. In this thesis, an ontology learning and mining model is proposed to acquire user pro¯les for personalised Web information gathering. The proposed compu- tational model emphasises the speci¯c is-a and part-of semantic relations in one computational model. The world knowledge and users' Local Instance Reposito- ries are used to attempt to discover and specify user background knowledge. From a world knowledge base, personalised ontologies are constructed by adopting au- tomatic or semi-automatic techniques to extract user interest concepts, focusing on user information needs. A multidimensional ontology mining method, Speci- ¯city and Exhaustivity, is also introduced in this thesis for analysing the user background knowledge discovered and speci¯ed in user personalised ontologies. The ontology learning and mining model is evaluated by comparing with human- based and state-of-the-art computational models in experiments, using a large, standard data set. The experimental results are promising for evaluation. The proposed ontology learning and mining model in this thesis helps to develop a better understanding of user pro¯le acquisition, thus providing better design of personalised Web information gathering systems. The contributions are increasingly signi¯cant, given both the rapid explosion of Web information in recent years and today's accessibility to the Internet and the full text world.

Effective pattern taxonomy mining in text documents

Many data mining techniques have been proposed for mining useful patterns in databases. However, how to effectively utilize discovered patterns is still an open research issue, especially in the domain of text mining. Most existing methods adopt term-based approaches. However, they all suffer from the problems of polysemy and synonymy. This paper presents an innovative technique, pattern taxonomy mining, to improve the effectiveness of using discovered patterns for finding useful information. Substantial experiments on RCV1 demonstrate that the proposed solution achieves encouraging performance.