Stable feature selection for clinical prediction: Exploiting ICD tree structure using Tree-Lasso

Modern healthcare is getting reshaped by growing Electronic Medical Records (EMR). Recently, these records have been shown of great value towards building clinical prediction models. In EMR data, patients' diseases and hospital interventions are captured through a set of diagnoses and procedures codes. These codes are usually represented in a tree form (e.g. ICD-10 tree) and the codes within a tree branch may be highly correlated. These codes can be used as features to build a prediction model and an appropriate feature selection can inform a clinician about important risk factors for a disease. Traditional feature selection methods (e.g. Information Gain, T-test, etc.) consider each variable independently and usually end up having a long feature list. Recently, Lasso and related l1-penalty based feature selection methods have become popular due to their joint feature selection property. However, Lasso is known to have problems of selecting one feature of many correlated features randomly. This hinders the clinicians to arrive at a stable feature set, which is crucial for clinical decision making process. In this paper, we solve this problem by using a recently proposed Tree-Lasso model. Since, the stability behavior of Tree-Lasso is not well understood, we study the stability behavior of Tree-Lasso and compare it with other feature selection methods. Using a synthetic and two real-world datasets (Cancer and Acute Myocardial Infarction), we show that Tree-Lasso based feature selection is significantly more stable than Lasso and comparable to other methods e.g. Information Gain, ReliefF and T-test. We further show that, using different types of classifiers such as logistic regression, naive Bayes, support vector machines, decision trees and Random Forest, the classification performance of Tree-Lasso is comparable to Lasso and better than other methods. Our result has implications in identifying stable risk factors for many healthcare problems and therefore can potentially assist clinical decision making for accurate medical prognosis.

An assessment of riparian restoration outcomes in two rural catchments in south-western Victoria: Focusing on tree and shrub species richness, structure and recruitment characteristics

Riparian ecosystems are among the most degraded systems in the landscape,and there has been substantial investment in their restoration. Consequently, monitoring restoration interventions offers opportunities to further develop the science of riparian restoration, particularly how to move from small-scale implementation to a broader landscape scale. Here, we report on a broad range of riparian revegetation projects in two regions of south-western Victoria, the Corangamite and Glenelg-Hopkins Catchment Management Areas. The objectives of restoration interventions in these regions have been stated quite broadly, for example, to reinstate terrestrial habitat and biodiversity, control erosion and improve water quality. This study reports on tree and shrub composition, structure and recruitment after restoration works compared with remnant vegetation found regionally. Within each catchment, a total of 57 sites from six subcatchments were identified, representing three age-classes: <4, 4–8 and >8–12 years after treatment, as well as untreated (control) sites. Treatments comprised fencing to exclude stock, spraying or slashing to reduce weed cover, followed by planting with tube stock. Across the six subcatchments, 12 reference (remnant) sites were used to provide a benchmark for species richness, structural and recruitment characteristics and to aid interpretation of the effects of the restoration intervention. Vegetation structure was well developed in the treated sites by 4–8 years after treatment. However, structural complexity was higher at remnant sites than at treated or untreated sites due to a higher richness of small shrubs. Tree and shrub recruitment occurred in all remnant sites and at 64% of sites treated >4 years ago. Most seedling recruitment at treatment sites was by Acacia spp. This assessment provides data on species richness, structure and recruitment characteristics following restoration interventions. Data from this study will contribute to longitudinal studies of vegetation processes in riparian landscapes of south-western Victoria.

Tree-based iterated local search for Markov random fields with applications in image analysis

The maximum a posteriori assignment for general structure Markov random fields is computationally intractable. In this paper, we exploit tree-based methods to efficiently address this problem. Our novel method, named Tree-based Iterated Local Search (T-ILS), takes advantage of the tractability of tree-structures embedded within MRFs to derive strong local search in an ILS framework. The method efficiently explores exponentially large neighborhoods using a limited memory without any requirement on the cost functions. We evaluate the T-ILS on a simulated Ising model and two real-world vision problems: stereo matching and image denoising. Experimental results demonstrate that our methods are competitive against state-of-the-art rivals with significant computational gain.

Graphene networks and their influence on free-volume properties of graphene-epoxidized natural rubber composites with a segregated structure: rheological and positron annihilation studies

Epoxidized natural rubber-graphene (ENR-GE) composites with segregated GE networks were successfully fabricated using the latex mixing combined in situ reduced technology. The rheological behavior and electrical conductivity of ENR-GE composites were investigated. At low frequencies, the storage modulus (G′) became frequency-independent suggesting a solid-like rheological behavior and the formation of GE networks. According to the percolation theory, the rheological threshold of ENR-GE composites was calculated to be 0.17 vol%, which was lower than the electrical threshold of 0.23 vol%. Both percolation thresholds depended on the evolution of the GE networks in the composites. At low GE concentrations (<0.17 vol%), GE existed as individual units, while a "polymer-bridged GE network" was constructed in the composites when GE concentrations exceeded 0.17 vol%. Finally, a "three-dimensional GE network" with percolation conductive paths was formed with a GE concentration of 0.23 vol%, where a remarkable increase in the conductivity of ENR-GE composites was observed. The effect of GE on the atom scale free-volume properties of composites was further studied by positron annihilation lifetime spectroscopy and positron age momentum correlation measurements. The motion of ENR chains was retarded by the geometric confinement of "GE networks", producing a high-density interfacial region in the vicinity of GE nanoplatelets, which led to a lower ortho-positronium lifetime intensity and smaller free-volume hole size.

In situ synthesis of natural rubber latex-supported gold nanoparticles for flexible SERS substrates

Natural rubber latex (NRL) from Hevea brasiliensis was used as a matrix to synthesize gold nanoparticles (AuNPs), leading to an organic-inorganic hybrid latex of NRL-supported AuNPs (AuNPs@NRL). The in situ and environmentally friendly preparation of AuNPs in an NRL matrix was developed by thermal treatment without using any other reducing agents or stabilizers because natural rubber particles and non-rubber components present in serum can serve as supporters for the synthesized AuNPs. As a result, the nanosized and well-dispersed AuNPs not only are decorated on the surface of natural rubber particles, but also can be found in the serum of NRL. The size of the AuNPs presented in NRL matrix can be controlled by adjusting the concentration of NRL. Furthermore, the flexible surface-enhanced Raman scattering (SERS) substrates made from the AuNPs@NRL through vacuum filtration presented good enhancement of the Raman probe molecule of 4-mercaptopyridine and outstanding SERS reproducibility. The capability of synthesizing the bio-supported nanohybrid latex provides a novel green and simple approach for the fabrication of flexible and effective SERS substrates.

Clustering and visualization of failure modes using an evolving tree

Despite the popularity of Failure Mode and Effect Analysis (FMEA) in a wide range of industries, two well-known shortcomings are the complexity of the FMEA worksheet and its intricacy of use. To the best of our knowledge, the use of computation techniques for solving the aforementioned shortcomings is limited. As such, the idea of clustering and visualization pertaining to the failure modes in FMEA is proposed in this paper. A neural network visualization model with an incremental learning feature, i.e., the evolving tree (ETree), is adopted to allow the failure modes in FMEA to be clustered and visualized as a tree structure. In addition, the ideas of risk interval and risk ordering for different groups of failure modes are proposed to allow the failure modes to be ordered, analyzed, and evaluated in groups. The main advantages of the proposed method lie in its ability to transform failure modes in a complex FMEA worksheet to a tree structure for better visualization, while maintaining the risk evaluation and ordering features. It can be applied to the conventional FMEA methodology without requiring additional information or data. A real world case study in the edible bird nest industry in Sarawak (Borneo Island) is used to evaluate the usefulness of the proposed method. The experiments show that the failure modes in FMEA can be effectively visualized through the tree structure. A discussion with FMEA users engaged in the case study indicates that such visualization is helpful in comprehending and analyzing the respective failure modes, as compared with those in an FMEA table. The resulting tree structure, together with risk interval and risk ordering, provides a quick and easily understandable framework to elucidate important information from complex FMEA forms; therefore facilitating the decision-making tasks by FMEA users. The significance of this study is twofold, viz., the use of a computational visualization approach to tackling two well-known shortcomings of FMEA; and the use of ETree as an effective neural network learning paradigm to facilitate FMEA implementations. These findings aim to spearhead the potential adoption of FMEA as a useful and usable risk evaluation and management tool by the wider community.

Searching Baxter's URDF robot joint and link tree for active serial chains

Teleoperation is integral to society's uptake of modern robotic systems. Given the wide array of readily available robots, ranging from simple mobile platforms and UAVs to advanced humanoid robots such as ASIMO and PR2, teleoperation is required in many different forms. The recent advances in virtual reality systems, interactive input controls and even haptic devices facilitate a wide range of new approaches to teleoperation control. This paper considers a dynamic user interface for improving the operator's ability to teleoperate heterogeneous robotic systems in dynamic and challenging environments. In order to achieve the proposed dynamic user interface the robot(s) comprising the heterogeneous robotic system and their active components need to be categorized. The recent uptake of ROS means that many robots are now represented within the standardized Unified Robot Descriptive Format (URDF), and this paper proposes a method for searching the URDF for active serial chains in individual robot systems. Results demonstrate the ability of the approach to determine active serial chains and associated kinematic information for the Baxter torso robot.

Experimental investigation on bonding properties of reactive liquid rubber epoxy in CFRP retrofitted concretet members

The load bearing capacity of aging reinforced concrete structures, such as bridges, is increasingly extended with the use of Carbon Fibre Reinforced Polymer (CFRP). Premature failure, which is attributed to the rigid behaviour of the bonding agent (epoxy resin) and the high stresses at the interface region, can occur because of the debonding of CFRP sheets from host surfaces. To overcome the debonding issue, the epoxy resin is modified by different reactive liquid polymers to improve its toughness, flexibility, adhesion, and impact resistance. This study reports the usage of two reactive liquid polymers, namely, liquid Carboxyl-Terminated Butadiene-Acrylonitrile (CTBN) and liquid Amine-Terminated Butadiene-Acrylonitrile (ATBN), to improve the mechanical properties of the commercially available MBrace saturant resin when added to a ratio of 100:30 by weight. The neat and modified epoxies were analysed using the Dynamic Mechanical Thermal Analysis (DMTA) to determine and compare the storage modulus and glass transition temperatures of these materials. Moreover, the bonding strength of neat and modified epoxies was evaluated through single-lap shear tests on CFRP sheets bonded to concrete prisms. The results indicate that the modified resins exhibited improved ductility and toughness and became reasonably flexible compared with the neat epoxy resin. The improved properties will help delay the premature debonding failure in CFRP retrofitted concrete members.

LC–MS and GC–MS metabolite profiling of nickel(II) complexes in the latex of the nickel-hyperaccumulating tree Sebertia acuminata and identification of methylated aldaric acid as a new nickel(II) ligand

Targeted liquid chromatography–mass spectrometry (LC–MS) technology using size exclusion chromatography and metabolite profiling based on gas chromatography–mass spectrometry (GC–MS) were used to study the nickel-rich latex of the hyperaccumulating tree Sebertia acuminata. More than 120 compounds were detected, 57 of these were subsequently identified. A methylated aldaric acid (2,4,5-trihydroxy-3-methoxy-1,6-hexan-dioic acid) was identified for the first time in biological extracts and its structure was confirmed by 1D and 2D nuclear magnetic resonance (NMR) spectroscopy. After citric acid, it appears to be one of the most abundant small organic molecules present in the latex studied. Nickel(II) complexes of stoichiometry NiII:acid = 1:2 were detected for these two acids as well as for malic, itaconic, erythronic, galacturonic, tartaric, aconitic and saccharic acids. These results provide further evidence that organic acids may play an important role in the transport and possibly in the storage of metal ions in hyperaccumulating plants.