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.

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.

Simulation of dynamic recrystallization using random grid cellular automata

Computer simulation is a powerful tool to predict microstructure and its evolution in dynamic and post-dynamic recrystallization. CAFE proposed as an appropriate approach by combining finite element (FE) method and cellular automata (CA) for recrystallization simulation. In the current study, a random grid cellular automaton (CA), as micro-scale model, based on finite element (FE), as macro-scale method, has been used to study initial and evolving microstructural features; including nuclei densities, dislocation densities, grain size and grain boundary movement during dynamic recrystallization in a C-Mn steel. An optimized relation has been established between mechanical variables and evolving microstructure features during recrystallization and grain growth. In this model, the microstructure is defined as cells located within grains and grain boundaries while dislocations are randomly dispersed throughout microstructure. Changes of dislocation density during deformation are described considering hardening, recovery and recrystallization. Recrystallization is assumed to initiate near grain boundaries and nucleation rate was considered constant (site-saturated condition). The model produced a mathematical formulation which captured the initial and evolving microstructural entities and linked their effects to measurable macroscopic variables (e.g. stress).

Face classification by a random forest

This paper presents a random forest-based face image classification method. The random forest is an ensemble learning method that grows many classification trees. Each tree gives a classification. The forest selects the classification that has the most votes. Three experiments are performed. The random forest-based method together with several existing approaches are trained and evaluated. The experimental results are presented and discussed.

Monetary models of exchange rates and the random walk

Three alternative monetary models of exchange rate are tested using data on the Italian lira - US doIIar exchange rate. II is shown that up to the early 1990s these economic models perform better than the random walk model in out-of-sample forecasts.

Sense of community in science fiction random, part 1: Understanding sense of community in an international community of interest

Within the discipline of community psychology there is debate as to the dimensions underlying the construct psychological sense of community (PSOC). One of the few theoretical discussions is that of McMillan and Chavis (1986), who hypothesized four dimensions: Belonging; Fulfillment of Needs; Influence; and Shared Connections. Discussion has also emerged in the literature regarding the role of identification within PSOC. It has been suggested that differences in PSOC may be understood in terms of the degree to which members identify with their community (Fisher & Sonn, 1999). However, few studies have explored the place of identification in PSOC. In addition, while PSOC has been applied to both communities of interest and geographical communities, little research has looked in depth at PSOC within communities of interest. The current study therefore explored PSOC in science fiction fandom, a community of interest with membership from all over the world, by means of a questionnaire distributed at an international science fiction convention (N = 359). In an endeavor to clarify the underlying dimensions of PSOC, the questionnaire included several measures of PSOC, and measures of identification with the community. Results showed that science fiction fandom reported high levels of PSOC. Support emerged for McMillan and Chavis' (1986) four dimensions of PSOC, with the addition of a fifth dimension, that of Conscious Identification. These results, and implications for PSOC research, are discussed.

Methodology for extracting quasi-random samples from world wide web domains

The purpose of this paper is to describe a process for sampling specific domain name zones on the World Wide Web. Because of the size of the Web, sampling strategies must be employed in order to effectively model and study the Web business environment.  This paper discusses Various efforts employed to sample the Web, which ranged from random generation of Internet Protocol Addresses and domain names, to the process finally
employed to create descriptive models of the dot-com domain name zones. The paper suggests that sampling the Web Top Level Domains offers a reasonable alternative for business researchers because it requires only familiarity with the use of the simple Web utilities such as File Transfer
Protocols to obtain initial domain name listings.