Text mining with semantic annotation : using enriched text representation for entity-oriented retrieval, semantic relation identification and text clustering

This project is a step forward in the study of text mining where enhanced text representation with semantic information plays a significant role. It develops effective methods of entity-oriented retrieval, semantic relation identification and text clustering utilizing semantically annotated data. These methods are based on enriched text representation generated by introducing semantic information extracted from Wikipedia into the input text data. The proposed methods are evaluated against several start-of-art benchmarking methods on real-life data-sets. In particular, this thesis improves the performance of entity-oriented retrieval, identifies different lexical forms for an entity relation and handles clustering documents with multiple feature spaces.

A case study of utilising concept knowledge in a topic specific document collection

The use of ‘topic’ concepts has shown improved search performance, given a query, by bringing together relevant documents which use different terms to describe a higher level concept. In this paper, we propose a method for discovering and utilizing concepts in indexing and search for a domain specific document collection being utilized in industry. This approach differs from others in that we only collect focused concepts to build the concept space and that instead of turning a user’s query into a concept based query, we experiment with different techniques of combining the original query with a concept query. We apply the proposed approach to a real-world document collection and the results show that in this scenario the use of concept knowledge at index and search can improve the relevancy of results.

Robust clustering of multi-type relational data via a heterogeneous manifold ensemble

High-Order Co-Clustering (HOCC) methods have attracted high attention in recent years because of their ability to cluster multiple types of objects simultaneously using all available information. During the clustering process, HOCC methods exploit object co-occurrence information, i.e., inter-type relationships amongst different types of objects as well as object affinity information, i.e., intra-type relationships amongst the same types of objects. However, it is difficult to learn accurate intra-type relationships in the presence of noise and outliers. Existing HOCC methods consider the p nearest neighbours based on Euclidean distance for the intra-type relationships, which leads to incomplete and inaccurate intra-type relationships. In this paper, we propose a novel HOCC method that incorporates multiple subspace learning with a heterogeneous manifold ensemble to learn complete and accurate intra-type relationships. Multiple subspace learning reconstructs the similarity between any pair of objects that belong to the same subspace. The heterogeneous manifold ensemble is created based on two-types of intra-type relationships learnt using p-nearest-neighbour graph and multiple subspaces learning. Moreover, in order to make sure the robustness of clustering process, we introduce a sparse error matrix into matrix decomposition and develop a novel iterative algorithm. Empirical experiments show that the proposed method achieves improved results over the state-of-art HOCC methods for FScore and NMI.

Interventions to control nosocomial infections : study designs and statistical issues

There is a wide range of potential study designs for intervention studies to decrease nosocomial infections in hospitals. The analysis is complex due to competing events, clustering, multiple timescales and time-dependent period and intervention variables. This review considers the popular pre-post quasi-experimental design and compares it with randomized designs. Randomization can be done in several ways: randomization of the cluster [intensive care unit (ICU) or hospital] in a parallel design; randomization of the sequence in a cross-over design; and randomization of the time of intervention in a stepped-wedge design. We introduce each design in the context of nosocomial infections and discuss the designs with respect to the following key points: bias, control for nonintervention factors, and generalizability. Statistical issues are discussed. A pre-post-intervention design is often the only choice that will be informative for a retrospective analysis of an outbreak setting. It can be seen as a pilot study with further, more rigorous designs needed to establish causality. To yield internally valid results, randomization is needed. Generally, the first choice in terms of the internal validity should be a parallel cluster randomized trial. However, generalizability might be stronger in a stepped-wedge design because a wider range of ICU clinicians may be convinced to participate, especially if there are pilot studies with promising results. For analysis, the use of extended competing risk models is recommended.

Assessment and application of clustering techniques to atmospheric particle number size distribution for the purpose of source apportionment

Long-term measurements of particle number size distribution (PNSD) produce a very large number of observations and their analysis requires an efficient approach in order to produce results in the least possible time and with maximum accuracy. Clustering techniques are a family of sophisticated methods which have been recently employed to analyse PNSD data, however, very little information is available comparing the performance of different clustering techniques on PNSD data. This study aims to apply several clustering techniques (i.e. K-means, PAM, CLARA and SOM) to PNSD data, in order to identify and apply the optimum technique to PNSD data measured at 25 sites across Brisbane, Australia. A new method, based on the Generalised Additive Model (GAM) with a basis of penalised B-splines, was proposed to parameterise the PNSD data and the temporal weight of each cluster was also estimated using the GAM. In addition, each cluster was associated with its possible source based on the results of this parameterisation, together with the characteristics of each cluster. The performances of four clustering techniques were compared using the Dunn index and Silhouette width validation values and the K-means technique was found to have the highest performance, with five clusters being the optimum. Therefore, five clusters were found within the data using the K-means technique. The diurnal occurrence of each cluster was used together with other air quality parameters, temporal trends and the physical properties of each cluster, in order to attribute each cluster to its source and origin. The five clusters were attributed to three major sources and origins, including regional background particles, photochemically induced nucleated particles and vehicle generated particles. Overall, clustering was found to be an effective technique for attributing each particle size spectra to its source and the GAM was suitable to parameterise the PNSD data. These two techniques can help researchers immensely in analysing PNSD data for characterisation and source apportionment purposes.

Topical pattern based document modelling and relevance ranking

For traditional information filtering (IF) models, it is often assumed that the documents in one collection are only related to one topic. However, in reality users’ interests can be diverse and the documents in the collection often involve multiple topics. Topic modelling was proposed to generate statistical models to represent multiple topics in a collection of documents, but in a topic model, topics are represented by distributions over words which are limited to distinctively represent the semantics of topics. Patterns are always thought to be more discriminative than single terms and are able to reveal the inner relations between words. This paper proposes a novel information filtering model, Significant matched Pattern-based Topic Model (SPBTM). The SPBTM represents user information needs in terms of multiple topics and each topic is represented by patterns. More importantly, the patterns are organized into groups based on their statistical and taxonomic features, from which the more representative patterns, called Significant Matched Patterns, can be identified and used to estimate the document relevance. Experiments on benchmark data sets demonstrate that the SPBTM significantly outperforms the state-of-the-art models.

Performative digital asset management: To propose a framework and proof of concept model that effectively enables researchers to document, archive and curate their non-traditional research data

This cross disciplinary study was conducted as two research and development projects. The outcome is a multimodal and dynamic chronicle, which incorporates the tracking of spatial, temporal and visual elements of performative practice-led and design-led research journeys. The distilled model provides a strong new approach to demonstrate rigour in non-traditional research outputs including provenance and an 'augmented web of facticity'.

Complete-linkage clustering for voice activity detection in audio and visual speech

We propose a novel technique for conducting robust voice activity detection (VAD) in high-noise recordings. We use Gaussian mixture modeling (GMM) to train two generic models; speech and non-speech. We then score smaller segments of a given (unseen) recording against each of these GMMs to obtain two respective likelihood scores for each segment. These scores are used to compute a dissimilarity measure between pairs of segments and to carry out complete-linkage clustering of the segments into speech and non-speech clusters. We compare the accuracy of our method against state-of-the-art and standardised VAD techniques to demonstrate an absolute improvement of 15% in half-total error rate (HTER) over the best performing baseline system and across the QUT-NOISE-TIMIT database. We then apply our approach to the Audio-Visual Database of American English (AVDBAE) to demonstrate the performance of our algorithm in using visual, audio-visual or a proposed fusion of these features.

Hierarchical clustering of the genetic connectivity matrix reveals the network topology of gene action on brain microstructure: An N=531 twin study

Genetic correlation (rg) analysis determines how much of the correlation between two measures is due to common genetic influences. In an analysis of 4 Tesla diffusion tensor images (DTI) from 531 healthy young adult twins and their siblings, we generalized the concept of genetic correlation to determine common genetic influences on white matter integrity, measured by fractional anisotropy (FA), at all points of the brain, yielding an NxN genetic correlation matrix rg(x,y) between FA values at all pairs of voxels in the brain. With hierarchical clustering, we identified brain regions with relatively homogeneous genetic determinants, to boost the power to identify causal single nucleotide polymorphisms (SNP). We applied genome-wide association (GWA) to assess associations between 529,497 SNPs and FA in clusters defined by hubs of the clustered genetic correlation matrix. We identified a network of genes, with a scale-free topology, that influences white matter integrity over multiple brain regions.

Genetic clustering on the hippocampal surface for genome-wide association studies

Imaging genetics aims to discover how variants in the human genome influence brain measures derived from images. Genome-wide association scans (GWAS) can screen the genome for common differences in our DNA that relate to brain measures. In small samples, GWAS has low power as individual gene effects are weak and one must also correct for multiple comparisons across the genome and the image. Here we extend recent work on genetic clustering of images, to analyze surface-based models of anatomy using GWAS. We performed spherical harmonic analysis of hippocampal surfaces, automatically extracted from brain MRI scans of 1254 subjects. We clustered hippocampal surface regions with common genetic influences by examining genetic correlations (r(g)) between the normalized deformation values at all pairs of surface points. Using genetic correlations to cluster surface measures, we were able to boost effect sizes for genetic associations, compared to clustering with traditional phenotypic correlations using Pearson's r.

Automatic clustering of white matter fibers in brain diffusion mri with an application to genetics

To understand factors that affect brain connectivity and integrity, it is beneficial to automatically cluster white matter (WM) fibers into anatomically recognizable tracts. Whole brain tractography, based on diffusion-weighted MRI, generates vast sets of fibers throughout the brain; clustering them into consistent and recognizable bundles can be difficult as there are wide individual variations in the trajectory and shape of WM pathways. Here we introduce a novel automated tract clustering algorithm based on label fusion - a concept from traditional intensity-based segmentation. Streamline tractography generates many incorrect fibers, so our top-down approach extracts tracts consistent with known anatomy, by mapping multiple hand-labeled atlases into a new dataset. We fuse clustering results from different atlases, using a mean distance fusion scheme. We reliably extracted the major tracts from 105-gradient high angular resolution diffusion images (HARDI) of 198 young normal twins. To compute population statistics, we use a pointwise correspondence method to match, compare, and average WM tracts across subjects. We illustrate our method in a genetic study of white matter tract heritability in twins.