A maximal frequent itemset approach for web document clustering

To efficiently and yet accurately cluster Web documents is of great interests to Web users and is a key component of the searching accuracy of a Web search engine. To achieve this, this paper introduces a new approach for the clustering of Web documents, which is called maximal frequent itemset (MFI) approach. Iterative clustering algorithms, such as K-means and expectation-maximization (EM), are sensitive to their initial conditions. MFI approach firstly locates the center points of high density clusters precisely. These center points then are used as initial points for the K-means algorithm. Our experimental results tested on 3 Web document sets show that our MFI approach outperforms the other methods we compared in most cases, particularly in the case of large number of categories in Web document sets.

An immune network approach for web document clustering

The human immune system provides inspiration for solving a wide range of innovative problems. In this paper, we propse an immune network based approach for web document clustering. All the immune cells in the network competitively recognize the antigens (web documents) which are presented to the network one by one. The interaction between immune cells and an antigen leads to an augment of the network through the clonal selection and somatic mutation of the stimulated immune cells, while the interaction among immune cells results in a network compression. The structure of the immune network is well maintained by learning and self-regularity. We use a public web document data set to test the effectiveness of our method and compare it with other approaches. The experimental results demonstrate that the most striking advantage of immune-based data clustering is its adaptation in dynamic environment and the capability of finding new clusters automatically.

Preparing construction project managers to document professional growth

Deakin University’s School of Architecture and Building is renowned for producing graduates who possess relevant attributes that make them job ready for the building and construction industry. Graduate destination surveys indicate that in the last eight (8) years, 100% of all Infrastructure Logistics (Construction and Facility Management) course graduates found relevant employment. This success is a direct result of a curriculum that is responsive to industry needs alongside educational methodology that focuses on excellent teaching and research while seeking new ways of developing and delivering courses.

The Infrastructure Logistics course prepares graduates to successfully compete in today’s global job market, and allows them to showcase relevant knowledge and skills that are critical in seeking and sustaining employment. Traditionally, tailored resumes served this purpose; however, in many professional fields, professional portfolios are now becoming a more desirable way of providing a summary of relevant attributes alongside evidence of professional abilities.

Sustaining employment, appraisals, and applying for a promotion are often subject to adequate evidence of professional standards and growth. Professional bodies require records of contribution to Continuing Professional Development (CPD) schemes; and accrediting organisations require professionals applying for professional registration to provide documented evidence of their relevant experience and abilities. The Australian Institute of Project Management (AIPM 2007) requires candidates wanting to become Registered Project Managers (RegPM) to demonstrate their current work-based experience and competencies.

This paper reports on a teaching strategy adopted in the Project Management (PM) stream, offered as part of Infrastructure Logistic courses. The teaching strategy is based on a combination of constructivism theory of learning, problem and project based learning, and active learning. The strategy requires systematic reflection and conscious creation of documented evidence of PM attributes and competences in the form of a portfolio.

Preliminary results of action research monitoring the effectiveness of systematic reflection indicate that students respond very positively to the idea of professional journals and professional portfolios. Preliminary results also indicate that students accept reflection and conscious documentation of their achievements as an integral part of their study and future practice.

Impact of image organizations on multimedia document retrieval

In this paper we compare ranking effectiveness of heterogeneous multimedia document retrieval when different image organizations are used for formulating queries. The quality of image queries depends on the organization of images used to make queries which in turn significantly impacts retrieval precision. CBIR (content based information retrieval) needs an effective and efficient organization of images including user interface which must be part of the configuration parameters of image retrieval research.

Staking a truth claim : Constructing a document in two laws

The documentary 'Two Laws' constitutes a legal document in support of the Borroloola claim to their land and contributes to the decolonisation of the images of Aboriginal Australia, which have circulated within ethnographic cinema, television journalism and fiction film. The 'two laws' of the film's title refer to white law and 'the Law', the system which regulates Borroloola social interactions and relationships with the land.

Text document clustering with metric learning

One reason for semi-supervised clustering fail to deliver satisfactory performance in document clustering is that the transformed optimization problem could have many candidate solutions, but existing methods provide no mechanism to select a suitable one from all those candidates. This paper alleviates this problem by posing the same task as a soft-constrained optimization problem, and introduces the salient degree measure as an information guide to control the searching of an optimal solution. Experimental results show the effectiveness of the proposed method in the improvement of the performance, especially when the amount of priori domain knowledge is limited.

Integrating instance-level and attribute-level knowledge into document clustering

In this paper, we present a document clustering framework incorporating instance-level knowledge in the form of pairwise constraints and attribute-level knowledge in the form of keyphrases. Firstly, we initialize weights based on metric learning with pairwise constraints, then simultaneously learn two kinds of knowledge by combining the distance-based and the constraint-based approaches, finally evaluate and select clustering result based on the degree of users’ satisfaction. The experimental results demonstrate the effectiveness and potential of the proposed method.

Probabilistic models over ordered partitions with applications in document ranking and collaborative filtering

Ranking is an important task for handling a large amount of content. Ideally, training data for supervised ranking would include a complete rank of documents (or other objects such as images or videos) for a particular query. However, this is only possible for small sets of documents. In practice, one often resorts to document rating, in that a subset of documents is assigned with a small number indicating the degree of relevance. This poses a general problem of modelling and learning rank data with ties. In this paper, we propose a probabilistic generative model, that models the process as permutations over partitions. This results in super-exponential combinatorial state space with unknown numbers of partitions and unknown ordering among them. We approach the problem from the discrete choice theory, where subsets are chosen in a stagewise manner, reducing the state space per each stage significantly. Further, we show that with suitable parameterisation, we can still learn the models in linear time. We evaluate the proposed models on two application areas: (i) document ranking with the data from the recently held Yahoo! challenge, and (ii) collaborative filtering with movie data. The results demonstrate that the models are competitive against well-known rivals.