Classification of time series by shapelet transformation

Time-series classification (TSC) problems present a specific challenge for classification algorithms: how to measure similarity between series. A shapelet is a time-series subsequence that allows for TSC based on local, phase-independent similarity in shape. Shapelet-based classification uses the similarity between a shapelet and a series as a discriminatory feature. One benefit of the shapelet approach is that shapelets are comprehensible, and can offer insight into the problem domain. The original shapelet-based classifier embeds the shapelet-discovery algorithm in a decision tree, and uses information gain to assess the quality of candidates, finding a new shapelet at each node of the tree through an enumerative search. Subsequent research has focused mainly on techniques to speed up the search. We examine how best to use the shapelet primitive to construct classifiers. We propose a single-scan shapelet algorithm that finds the best $k$ shapelets, which are used to produce a transformed dataset, where each of the $k$ features represent the distance between a time series and a shapelet. The primary advantages over the embedded approach are that the transformed data can be used in conjunction with any classifier, and that there is no recursive search for shapelets. We demonstrate that the transformed data, in conjunction with more complex classifiers, gives greater accuracy than the embedded shapelet tree. We also evaluate three similarity measures that produce equivalent results to information gain in less time. Finally, we show that by conducting post-transform clustering of shapelets, we can enhance the interpretability of the transformed data. We conduct our experiments on 29 datasets: 17 from the UCR repository, and 12 we provide ourselves.     

Gamma linolenic acid (GLA) content of encapsulated evening primrose oil products

The fatty acid composition of 16 brands of evening primrose oil (EPO) capsules was determined by capillary gas chromatography. Fourteen of these EPO brands contained γ-linolenic acid (GLA) levels between 7% and 10% (mean, 8.7; range, 1.9–10.5%) and there was generally good agreement between the level of GLA claimed by the manufacturer and the level determined by analysis. Low levels of the monoenes 22∶1 and 24∶1 found in some brands may indicate contamination of EPO with borage oil.     

Thermal expansion — An empirical correlation

The average thermal expansion coefficient (α_L) of a large variety of cubic and close-packed materials is shown to depend upon the coincidence of preferred orbital extension and site symmetry as well as melting point (Mp in °C). The product $\text{α}{}_{\mathrm{<mtext>L</mtext>}}\text{M}{}_{\mathrm{<mtext>p</mtext>}}\text{? 0.016}$ tends to hold for close-packed structures and $\text{α}{}_{\mathrm{<mtext>L</mtext>}}\text{M}{}_{\mathrm{<mtext>p</mtext>}}\text{? 0.027}$ for rectilinear arrays. Deviations occur when orbital extension tends to sp³ symmetry for an ion in an octahedral site or the site symmetry is reduced to tetrahedral or planar.     

Optimization of hydrogen stations in Florida using the Flow-Refueling Location Model

This paper develops and applies a model that locates hydrogen stations to refuel the maximum volume of vehicle flows. Inputs to the model include a road network with average speeds; the origin–destination flow volumes between each origin and destination; a maximum driving range between refueling stops; and the number of stations to build. The Flow-Refueling Location Model maximizes the flow volumes that can be refueled, measured either in number of trips or vehicle-miles traveled. Geographic Information Systems and heuristic algorithms are integrated in a spatial decision support system that researchers can use to develop data, enter assumptions, analyze scenarios, evaluate tradeoffs, and map results. For the Florida Hydrogen Initiative, we used this model to investigate strategies for rolling out an initial refueling infrastructure in Florida at two different scales of analysis: metropolitan Orlando and statewide. By analyzing a variety of scenarios at both scales of analysis, we identify a robust set of stations that perform well under a variety of assumptions, and develop a strategy for phasing in clustered and connecting stations in several stages or tiers.     

Adaptive finite element simulation of ventricular fibrillation dynamics

The paper communicates simulation results (computations and visualizations) for the dynamics of ventricular fibrillation caused by irregular excitation in the frame of the monodomain model with an action potential model due to Aliev–Panfilov for a human 3D geometry. The numerical solution of this challenging multiscale reaction–diffusion problem is attacked by algorithms which are fully adaptive in both space and time (code library KARDOS). The obtained results clearly demonstrate an accurate resolution of the cardiac potential during the excitation and the plateau phases (in the regular cycle) as well as after a reentrant excitation (in the irregular cycle). Supported by the DFG Research Center Matheon “Mathematics for key technologies” in Berlin.