Graph matching : fast candidate elimination using machine learning techniques

Graph matching is an important class of methods in pattern recognition. Typically, a graph representing an unknown pattern is matched with a database of models. If the database of model graphs is large, an additional factor in induced into the overall complexity of the matching process. Various techniques for reducing the influence of this additional factor have been described in the literature. In this paper we propose to extract simple features from a graph and use them to eliminate candidate graphs from the database. The most powerful set of features and a decision tree useful for candidate elimination are found by means of the C4.5 algorithm, which was originally proposed for inductive learning of classication rules. Experimental results are reported demonstrating that effcient candidate elimination can be achieved by the proposed procedure.

Preserving utility in social network graph anonymization

To protect from privacy disclosure, the social network graph is modified in order to hide the information that potentially be used to disclose person's identity. However, when the social network graph is changed, it is a great challenge to balance between the privacy gained and the loss of data utility. In this paper, we address this problem. We propose a new graph topological-based metric to improve utility preservation in social network graph anonymization. We compare the proposed approach with the amount-of-edge-change metric that popularly used in most of previous works. Experimental evaluation shows that our approach generates anonymized social network with improved utility preservation.

On the edge: how well do fire mitigation strategies work on the urban fringe?

The reduction of loss of lives and assets during bushfire is one of the primary aims or lire management agencies. Traditional fire mitigation strategies include strategic fire hreaks, static water points, management of ignition sources, rapid detection and local response, air attack, and fuel reduction burning. There have been few quantitative studies that assess the success or these strategies. We need to promote 'new' strategies more focused on human hehaviour and community preparedness.
DeJcndable space provides our best strategy for reducing losses during major bushfires. The size or the defendable space depends on the type of house to be defended, who is defending it, and the spatial context of the property. In the urban fi'inge, remnant vegetation on private property often has
high conservation values, and application of traditional mitigation strategies, as well as the vegetation modification required to achieve defendable space, may have significant impacts on conservation and biodiversity values.

Development of the sheared edge in the trimming of steel and light metal sheet, part 1 - experimental observations

Trimming experiments were conducted on sheet metals including two drawing steels, an aluminum alloy and a magnesium alloy, using a specially designed die in a mechanical press. The punch-die clearance was varied and data obtained on the rollover and burr height as a function of the clearance. Samples were also partially trimmed to examine crack initiation, the generation of the fracture surface profile and mechanism of burr formation. The results showed that while the burr height and rollover depth generally increased with increasing clearance for all examined materials, there were differences in the fracture surface profile shape, the burr shape, and the mechanism of burr formation, between the two steels and the two light alloys. The major cause of these differences appeared to be the rate of crack propagation through the sheet material.

Development of the sheared edge in the trimming of steel and light metal sheet, part 2 - mechanisms and modelling

The shearing behavior of a drawing-steel and aluminum alloy were investigated using hardness contours of partially deformed samples and a finite element model of the trimming process. Results showed that the stress and strain distributions within the work-piece were more strongly dependent on the punch penetration than the material properties of the work-piece. Differences in the final fracture surface profile and burr formation of the drawing-steel and aluminum alloy were a consequence of the shape of the stress and strain distribution when the crack in the sample became unstable, not when it was initiated. Results and existing literature suggest that a correlation may exist between the strain-rate sensitivity of the work-piece material and the burr mechanism and fracture surface profile of the trimmed part.

Influence of z-pinning on the buckling of composite laminates under edge-wise compression

Z-pinning is a newly developed technique to enhance the strength of composite laminates in the thickness direction. Recent experimental and theoretical studies have shown that z-pins significantly improve mode I and mode II fracture toughness. In practice, buckling accompanying delamination is a typical failure mode in laminated composite structures. For a complete understanding of the z-pinning technique towards improvements of the overall mechanical properties of laminated composites, a numerical model is developed in this paper to investigate the influence of z-pins on the buckling composite laminates with initial delaminations under edge-wise compression. The numerical results indicate that z-pinning can indeed effectively increase the compressive strength of the composite laminates provided that the initial imperfection is within a certain range. The magnitude of the improvement is consistent with available experimental data.

A study on reliability in graph discovery

A critical question in data mining is that can we always trust what discovered by a data mining system unconditionally? The answer is obviously not. If not, when can we trust the discovery then? What are the factors that affect the reliability of the discovery? How do they affect the reliability of the discovery? These are some interesting questions to be investigated.

In this paper we will firstly provide a definition and the measurements of reliability, and analyse the factors that affect the reliability. We then examine the impact of model complexity, weak links, varying sample sizes and the ability of different learners to the reliability of graphical model discovery. The experimental results reveal that (1) the larger sample size for the discovery, the higher reliability we will get; (2) the stronger a graph link is, the easier the discovery will be and thus the higher the reliability it can achieve; (3) the complexity of a graph also plays an important role in the discovery. The higher the complexity of a graph is, the more difficult to induce the graph and the lower reliability it would be.

Effect of proximity to the shelf edge on the diet of female Australian fur seals

The Australian fur seal Arctocephalus pusillus doriferus is a temperate latitude species with a breeding distribution restricted to Bass Strait, Australia. Recent studies of the foraging behaviour of female Australian fur seals indicated that they feed demersally in the shallow continental shelf waters, a behaviour that is in contrast to the epipelagic foraging of females of most other arctocephaline (Southern Hemisphere fur seals) species and akin to that observed in sea lions. These studies, however, were conducted at one colony (Kanowna Island) located in central northern Bass Strait, and it was suggested that the observed foraging behaviour may have been due to the distance of this colony from the continental shelf edge (180 km), making it inefficient to forage beyond it. Here, the diet of lactating Australian fur seals was compared between 2 colonies to test if differing proximity to the continental shelf edge resulted in differences in foraging behaviour. The 2 breeding colonies studied, Kanowna Island and The Skerries, were 180 and 25 km from the nearest shelf edge, respectively. We analysed a total of 917 scat samples collected at the 2 colonies between 1997 and 2001. From faecal analysis, 45 primarily demersal on-shelf species of fishes and cephalopods were identified. Only 4 species had a frequency of occurrence greater than 10%: redbait Emmelichthys nitidus, jack mackerel Trachurus sp., red rock cod Pseudophycis bachus, and Gould’s squid Nototodarus gouldi. No seasonal, annual or spatial differences were found between the 2 colonies, indicating that proximity to the shelf edge does not influence diet.

Polyhedral studies for minimum-span graph labelling with integer distance constraints

This paper studies the polytope of the minimum-span graph labelling problems with integer distance constraints (DC-MSGL). We first introduce a few classes of new valid inequalities for the DC-MSGL defined on general graphs and briefly discuss the separation problems of some of these inequalities. These are the initial steps of a branch-and-cut algorithm for solving the DC-MSGL. Following that, we present our polyhedral results on the dimension of the DC-MSGL polytope, and that some of the inequalities are facet defining, under reasonable conditions, for the polytope of the DC-MSGL on triangular graphs.

Synthesis of [n]polynorbornanes with differing edge substitution: a new class of regioselectively addressable framework

Polynorbornanes with differing edge functionality have been synthesised from the appropriate cyclobutene epoxides substituted with two, unlike, electron withdrawing groups. These latter compounds were prepared by the monohydrolysis of symmetric cyclobutene diesters and subsequent elaboration of the resulting carboxylic acid.