Itemset mining: A constraint programming perspective

The field of data mining has become accustomed to specifying constraints on patterns of interest. A large number of systems and techniques has been developed for solving such constraint-based mining problems, especially for mining itemsets. The approach taken in the field of data mining contrasts with the constraint programming principles developed within the artificial intelligence community. While most data mining research focuses on algorithmic issues and aims at developing highly optimized and scalable implementations that are tailored towards specific tasks, constraint programming employs a more declarative approach. The emphasis lies on developing high-level modeling languages and general solvers that specify what the problem is, rather than outlining how a solution should be computed, yet are powerful enough to be used across a wide variety of applications and application domains.This paper contributes a declarative constraint programming approach to data mining. More specifically, we show that it is possible to employ off-the-shelf constraint programming techniques for modeling and solving a wide variety of constraint-based itemset mining tasks, such as frequent, closed, discriminative, and cost-based itemset mining. In particular, we develop a basic constraint programming model for specifying frequent itemsets and show that this model can easily be extended to realize the other settings. This contrasts with typical procedural data mining systems where the underlying procedures need to be modified in order to accommodate new types of constraint, or novel combinations thereof. Even though the performance of state-of-the-art data mining systems outperforms that of the constraint programming approach on some standard tasks, we also show that there exist problems where the constraint programming approach leads to significant performance improvements over state-of-the-art methods in data mining and as well as to new insights into the underlying data mining problems. Many such insights can be obtained by relating the underlying search algorithms of data mining and constraint programming systems to one another. We discuss a number of interesting new research questions and challenges raised by the declarative constraint programming approach to data mining.     

Numerical Analysis of Coaxial Double Gate Schottky Barrier Carbon Nanotube Field Effect Transistors

Carbon nanotube field-effect transistors (CNTFETs) have been studied in recent years as a potential alternative to CMOS devices, because of the capability of ballistic transport. The ambipolar behavior of Schottky barrier CNTFETs limits the performance of these devices. A double gate design is proposed to suppress this behavior. In this structure the first gate located near the source contact controls carrier injection and the second gate located near the drain contact suppresses parasitic carrier injection. To avoid the ambipolar behavior it is necessary that the voltage of the second gate is higher or at least equal to the drain voltage. The behavior of these devices has been studied by solving the coupled Schrödinger-Poisson equation system. We investigated the effect of the second gate voltage on the performance of the device and finally the advantages and disadvantages of these options are discussed.     

A 2 MV Van de Graaff accelerator as a tool for planetary and impact physics research

Investigating the dynamical and physical properties of cosmic dust can reveal a great deal of information about both the dust and its many sources. Over recent years, several spacecraft (e.g., Cassini, Stardust, Galileo, and Ulysses) have successfully characterised interstellar, interplanetary, and circumplanetary dust using a variety of techniques, including in situ analyses and sample return. Charge, mass, and velocity measurements of the dust are performed either directly (induced charge signals) or indirectly (mass and velocity from impact ionisation signals or crater morphology) and constrain the dynamical parameters of the dust grains. Dust compositional information may be obtained via either time-of-flight mass spectrometry of the impact plasma or direct sample return. The accurate and reliable interpretation of collected spacecraft data requires a comprehensive programme of terrestrial instrument calibration. This process involves accelerating suitable solar system analogue dust particles to hypervelocity speeds in the laboratory, an activity performed at the Max Planck Institut fu?r Kernphysik in Heidelberg, Germany. Here, a 2 MV Van de Graaff accelerator electrostatically accelerates charged micron and submicron-sized dust particles to speeds up to 80 km s(-1). Recent advances in dust production and processing have allowed solar system analogue dust particles (silicates and other minerals) to be coated with a thin conductive shell, enabling them to be charged and accelerated. Refinements and upgrades to the beam line instrumentation and electronics now allow for the reliable selection of particles at velocities of 1-80 km s(-1) and with diameters of between 0.05 μm and 5 μm. This ability to select particles for subsequent impact studies based on their charges, masses, or velocities is provided by a particle selection unit (PSU). The PSU contains a field programmable gate array, capable of monitoring in real time the particles' speeds and charges, and is controlled remotely by a custom, platform independent, software package. The new control instrumentation and electronics, together with the wide range of accelerable particle types, allow the controlled investigation of hypervelocity impact phenomena across a hitherto unobtainable range of impact parameters.     

Evolution of Current Transport Models for Engineering Applications

An overview of models for the simulation of current transport in micro- and nanoelectronic devices within the framework of TCAD applications is presented. Starting from macroscopic transport models, currently discussed enhancements are specifically addressed. This comprises the inclusion of higher-order moments into the transport models, the incorporation of quantum correction and tunneling models up to dedicated quantum-mechanical simulators, and mixed approaches which are able to account for both, quantum interference and scattering. Specific TCAD requirements are discussed from an engineer’s perspective and an outlook on future research directions is given.     

Impact of amorphous and crystalline lactose on milk chocolate properties

Chocolate mass of low viscosity is preferred for most applications. Milk powder influences processing behaviour, flow properties and taste of milk chocolate. The project aimed to investigate influences of skim milk powders containing amorphous or crystalline lactose on flow properties after producing samples by roller milling and conching or alternatively by ball milling. For the first case, it was found that mass consistency before roller milling is strongly influenced by lactose type; producers must specify it and adapt initial mass fat content. Little impact on final products was found after processing milk powders at equilibrium moisture. If predried powders are used for reducing conching time, crystalline lactose leads to chocolate with slightly lower viscosity. At ball mill processing, crystalline lactose resulted in significantly lower viscosity, for example 15% at 40 s^?1; thus, for this process, it can be recommended to use special milk powders high in crystalline lactose content.     

Determination of nano-roughness of carbon fibers by atomic force microscopy

We present a novel approach to determine the surface roughness on varying scales using atomic force microscopy data. The key factor is to find a suitable background correction for the desired scale. Using the example of the surface of sized and unsized high-tenacity carbon fibers, we present an easy method to find backgrounds for widely varying scales and to evaluate respective topography and surface roughness with the same lateral resolution as the microscope itself. The analysis is done by subtracting a tunable background from the respective height data. By choosing an appropriate background to investigate the surface topography of a carbon fiber on a nm-scale, only small nano-structures with a width of around 20 nm remain after the background subtraction. Evaluating the mean roughness R _a of these nano-structures, sized carbon fibers show an overall value of around 0.1 nm while unsized carbon fibers a value of around 0.4 nm. Total background corrected height analysis shows an even distribution of these nano-structures along the fibrils of the unsized fibers, whereas for the sized fibers the nano-structures are not present. The presented method allows analysis and visualization of the distribution of nano-structures on a carbon fiber surface for the first time. This feature is used to visualize the distribution of the sizing and can further be used to investigate the influence of different production parameters on the fiber topography or to evaluate the contribution of mechanical interlocking to the interfacial strength.