Neural and wavelet network models for financial distress classification

This work analyzes the use of linear discriminant models, multi-layer perceptron neural networks and wavelet networks for corporate financial distress prediction. Although simple and easy to interpret, linear models require statistical assumptions that may be unrealistic. Neural networks are able to discriminate patterns that are not linearly separable, but the large number of parameters involved in a neural model often causes generalization problems. Wavelet networks are classification models that implement nonlinear discriminant surfaces as the superposition of dilated and translated versions of a single "mother wavelet" function. In this paper, an algorithm is proposed to select dilation and translation parameters that yield a wavelet network classifier with good parsimony characteristics. The models are compared in a case study involving failed and continuing British firms in the period 1997-2000. Problems associated with over-parameterized neural networks are illustrated and the Optimal Brain Damage pruning technique is employed to obtain a parsimonious neural model. The results, supported by a re-sampling study, show that both neural and wavelet networks may be a valid alternative to classical linear discriminant models.

Increased levels of microparticles originating from endothelial cells, platelets and erythrocytes in subjects with metabolic syndrome: relationship with oxidative stress

Background and aims The Metabolic Syndrome (MetS) is associated with increased cardiovascular risk. Circulating microparticles (MP) are involved in the pathogenesis of atherothrombotic disorders and are raised in individual with CVD. We measured their level and cellular origin in subjects with MetS and analyzed their associations with 1/anthropometric and biological parameters of MetS, 2/inflammation and oxidative stress markers. Methods and results Eighty-eight subjects with the MetS according to the NCEP-ATPIII definition were enrolled in a bicentric study and compared to 27 healthy controls. AnnexinV-positive MP (TMP), MP derived from platelets (PMP), erythrocytes (ErMP), endothelial cells (EMP), leukocytes (LMP) and granulocytes (PNMP) were determined by flow cytometry. MetS subjects had significantly higher counts/μl of TMP (730.6 ± 49.7 vs 352.8 ± 35.6), PMP (416.0 ± 43.8 vs 250.5 ± 23.5), ErMP (243.8 ± 22.1 vs 73.6 ± 19.6) and EMP (7.8 ± 0.8 vs 4.0 ± 1.0) compared with controls. LMP and PNMP were not statistically different between groups. Multivariate analysis demonstrated that each criterion for the MetS influenced the number of TMP. Waist girth was a significant determinant of PMP and EMP level and blood pressure was correlated with EMP level. Glycemia positively correlated with PMP level whereas dyslipidemia influenced EMP and ErMP levels. Interestingly, the oxidative stress markers, plasma glutathione peroxydase and urinary 8-iso-prostaglandin F2 α, independently influenced TMP and PMP levels whereas inflammatory markers did not, irrespective of MP type. Conclusion Increased levels of TMP, PMP, ErMP and EMP are associated with individual metabolic abnormalities of MetS and oxidative stress. Whether MP assessment may represent a marker for risk stratification or a target for pharmacological intervention deserves further investigation.

On combination of SMOTE and particle swarm optimization based radial basis function classifier for imbalanced problems

The combination of the synthetic minority oversampling technique (SMOTE) and the radial basis function (RBF) classifier is proposed to deal with classification for imbalanced two-class data. In order to enhance the significance of the small and specific region belonging to the positive class in the decision region, the SMOTE is applied to generate synthetic instances for the positive class to balance the training data set. Based on the over-sampled training data, the RBF classifier is constructed by applying the orthogonal forward selection procedure, in which the classifier structure and the parameters of RBF kernels are determined using a particle swarm optimization algorithm based on the criterion of minimizing the leave-one-out misclassification rate. The experimental results on both simulated and real imbalanced data sets are presented to demonstrate the effectiveness of our proposed algorithm.

A combined SMOTE and PSO based RBF classifier for two-class imbalanced problems

This contribution proposes a powerful technique for two-class imbalanced classification problems by combining the synthetic minority over-sampling technique (SMOTE) and the particle swarm optimisation (PSO) aided radial basis function (RBF) classifier. In order to enhance the significance of the small and specific region belonging to the positive class in the decision region, the SMOTE is applied to generate synthetic instances for the positive class to balance the training data set. Based on the over-sampled training data, the RBF classifier is constructed by applying the orthogonal forward selection procedure, in which the classifier's structure and the parameters of RBF kernels are determined using a PSO algorithm based on the criterion of minimising the leave-one-out misclassification rate. The experimental results obtained on a simulated imbalanced data set and three real imbalanced data sets are presented to demonstrate the effectiveness of our proposed algorithm.

Energy and protein interactions and their effect on nitrogen excretion in dairy cows

The principal driver of nitrogen (N) losses from the body including excretion and secretion in milk is N intake. However, other covariates may also play a role in modifying the partitioning of N. This study tests the hypothesis that N partitioning in dairy cows is affected by energy and protein interactions. A database containing 470 dairy cow observations was collated from calorimetry experiments. The data include N and energy parameters of the diet and N utilization by the animal. Univariate and multivariate meta-analyses that considered both within and between study effects were conducted to generate prediction equations based on N intake alone or with an energy component. The univariate models showed that there was a strong positive linear relationships between N intake and N excretion in faeces, urine and milk. The slopes were 0.28 faeces N, 0.38 urine N and 0.20 milk N. Multivariate model analysis did not improve the fit. Metabolizable energy intake had a significant positive effect on the amount of milk N in proportion to faeces and urine N, which is also supported by other studies. Another measure of energy considered as a covariate to N intake was diet quality or metabolizability (the concentration of metabolizable energy relative to gross energy of the diet). Diet quality also had a positive linear relationship with the proportion of milk N relative to N excreted in faeces and urine. Metabolizability had the largest effect on faeces N due to lower protein digestibility of low quality diets. Urine N was also affected by diet quality and the magnitude of the effect was higher than for milk N. This research shows that including a measure of diet quality as a covariate with N intake in a model of N execration can enhance our understanding of the effects of diet composition on N losses from dairy cows. The new prediction equations developed in this study could be used to monitor N losses from dairy systems.

Fulvalene titanium and zirconium complexes: synthesis and NMR study of phosphanido-, alkyl-, and alkynyl-derivatives. X-ray crystal structures of [{Ti(η5−C5H5)(μ−PPH2)}2{μ−(η5−C5H4−η5−C5H4)}] and [{Zr(η5−C5H5)(μ−C≡CSiMe3)}2{μ−(η5−C5H4−η5−C5H4)}]

Reaction of with one or two equivalents of LiPPh2 afforded the new phosphanidometal(III) complexes . Reaction of 2 with LiC≡CSiMe3 led to the diamagnetic zirconium(III) alkynyl derivative [{Zr(C5H5)(μ−C≡CSiMe3)}2(μ−η5−C5H4−η5−C5H4], 7. Alkylation of 6 with LiCH2CMe2Ph gave [{Zr(η5−C5H5)(CH2CMe2Ph)2}2{μ−(η5−C5H4)}], 8. A detailed NMR study of complexes 3 and 4 allowed the observation of the spectral behaviour of the eight different fulvalene protons through their coupling to the 31P nucleus. The fluxional behaviour of complex 7 was studied by dynamic DNMR, and kinetic parameters for the σ-π-conversion of the alkynyl ligand were determined. The molecular structures of complexes 3 and 7 were determined by X-ray diffraction methods.

Prediction of coagulation properties, titratable acidity and pH of bovine milk using mid-infrared spectroscopy

This study investigated the potential application of mid-infrared spectroscopy (MIR 4,000–900 cm−1) for the determination of milk coagulation properties (MCP), titratable acidity (TA), and pH in Brown Swiss milk samples (n = 1,064). Because MCP directly influence the efficiency of the cheese-making process, there is strong industrial interest in developing a rapid method for their assessment. Currently, the determination of MCP involves time-consuming laboratory-based measurements, and it is not feasible to carry out these measurements on the large numbers of milk samples associated with milk recording programs. Mid-infrared spectroscopy is an objective and nondestructive technique providing rapid real-time analysis of food compositional and quality parameters. Analysis of milk rennet coagulation time (RCT, min), curd firmness (a30, mm), TA (SH°/50 mL; SH° = Soxhlet-Henkel degree), and pH was carried out, and MIR data were recorded over the spectral range of 4,000 to 900 cm−1. Models were developed by partial least squares regression using untreated and pretreated spectra. The MCP, TA, and pH prediction models were improved by using the combined spectral ranges of 1,600 to 900 cm−1, 3,040 to 1,700 cm−1, and 4,000 to 3,470 cm−1. The root mean square errors of cross-validation for the developed models were 2.36 min (RCT, range 24.9 min), 6.86 mm (a30, range 58 mm), 0.25 SH°/50 mL (TA, range 3.58 SH°/50 mL), and 0.07 (pH, range 1.15). The most successfully predicted attributes were TA, RCT, and pH. The model for the prediction of TA provided approximate prediction (R2 = 0.66), whereas the predictive models developed for RCT and pH could discriminate between high and low values (R2 = 0.59 to 0.62). It was concluded that, although the models require further development to improve their accuracy before their application in industry, MIR spectroscopy has potential application for the assessment of RCT, TA, and pH during routine milk analysis in the dairy industry. The implementation of such models could be a means of improving MCP through phenotypic-based selection programs and to amend milk payment systems to incorporate MCP into their payment criteria.

Application of near and mid-infrared spectroscopy to determine cheese quality and authenticity

This paper reviews the current state of development of both near-infrared (NIR) and mid-infrared (MIR) spectroscopic techniques for process monitoring, quality control, and authenticity determination in cheese processing. Infrared spectroscopy has been identified as an ideal process analytical technology tool, and recent publications have demonstrated the potential of both NIR and MIR spectroscopy, coupled with chemometric techniques, for monitoring coagulation, syneresis, and ripening as well as determination of authenticity, composition, sensory, and rheological parameters. Recent research is reviewed and compared on the basis of experimental design, spectroscopic and chemometric methods employed to assess the potential of infrared spectroscopy as a technology for improving process control and quality in cheese manufacture. Emerging research areas for these technologies, such as cheese authenticity and food chain traceability, are also discussed.

Survival, reproduction and population growth of the bee pollinator, Osmia rufa (Hymenoptera: Megachilidae), along gradients of heavy metal pollution

1. Bees are one of the most important groups of pollinators in the temperate zone. Although heavy metal pollution is recognised to be a problem affecting large parts of the European Union, we currently lack insights into the effects of heavy metals on wild bee survival and reproduction. 2. We investigated the impact of heavy metal pollution on the wild bee Osmia rufa (Hymenoptera: Megachilidae) by comparing their survival, reproduction and population dynamics along two independent gradients of heavy metal pollution, one in Poland and the other in the United Kingdom. We used trap nests to evaluate the response of fitness and survival parameters of O. rufa. To quantify the levels of pollution, we directly measured the heavy metal concentration in provisions collected by O. rufa. 3. We found that with increasing heavy metal concentration, there was a steady decrease in number of brood cells constructed by females and an increase in the proportion of dead offspring. In the most polluted site, there were typically 3–4 cells per female with 50–60% dead offspring, whereas in unpolluted sites there were 8 to 10 cells per female and only 10–30% dead offspring. Moreover, the bee population growth rate (R0) decreased along the heavy metal pollution gradients. In unpolluted sites, R0 was above 1, whereas in contaminated sites, the values tended to be below 1. 4. Our findings reveal a negative relationship between heavy metal pollution and several fitness parameters of the wild bee O. rufa, and highlight a mechanism whereby the detrimental effects of heavy metal pollution may severely impact wild bee communities.

The surface urban energy and water balance scheme (SUEWS): evaluation in Los Angeles and Vancouver

An urban energy and water balance model is presented which uses a small number of commonly measured meteorological variables and information about the surface cover. Rates of evaporation-interception for a single layer with multiple surface types (paved, buildings, coniferous trees and/or shrubs, deciduous trees and/or shrubs, irrigated grass, non-irrigated grass and water) are calculated. Below each surface type, except water, there is a single soil layer. At each time step the moisture state of each surface is calculated. Horizontal water movements at the surface and in the soil are incorporated. Particular attention is given to the surface conductance used to model evaporation and its parameters. The model is tested against direct flux measurements carried out over a number of years in Vancouver, Canada and Los Angeles, USA. At all measurement sites the model is able to simulate the net all-wave radiation and turbulent sensible and latent heat well (RMSE = 25–47 W m−2, 30–64 and 20–56 W m−2, respectively). The model reproduces the diurnal cycle of the turbulent fluxes but typically underestimates latent heat flux and overestimates sensible heat flux in the day time. The model tracks measured surface wetness and simulates the variations in soil moisture content. It is able to respond correctly to short-term events as well as annual changes. The largest uncertainty relates to the determination of surface conductance. The model has the potential be used for multiple applications; for example, to predict effects of regulation on urban water use, landscaping and planning scenarios, or to assess climate mitigation strategies.

Circulating insulin-like factor 3 (INSL3) in healthy and infertile women

STUDY QUESTION: How does insulin-like factor 3 (INSL3) concentration in blood vary across the menstrual cycle in women? SUMMARY ANSWER: INSL3 is secreted by the theca interna cells of growing antral follicles and is phasic in its expression. WHAT IS KNOWN ALREADY: The relaxin-like hormone INSL3 is known to be expressed in follicles of several mammal species, and was recently shown in cows to be specifically secreted into the bloodstream by growing antral follicles, corresponding to follicular waves. In males INSL3 is known to be acutely independent of the hormones of the hypothalamic-pituitary-gonadal axis, suggesting that in women INSL3 might be a novel biomarker for antral follicle recruitment and development. STUDY DESIGN, SIZE, DURATION: Two cohorts of women were studied. First, 18 healthy women of reproductive age were followed longitudinally for one and a half cycles, with blood sampling and hormone measurement every 2-3 days. A second cohort comprised a cross-sectional study of 909 women attending an infertility clinic, with a single blood sample taken at entry, together with other clinical and hormonal parameters. PARTICIPANTS/MATERIALS, SETTING, METHODS: Blood samples from both retrospective cohorts were analyzed for INSL3 using a highly sensitive time-resolved fluorescent immunoassay, and data were analyzed in comparison with other clinical and hormonal parameters. MAIN RESULT AND THE ROLE OF CHANCE: For young healthy women of reproductive age, we showed a phasic expression of INSL3 corresponding to antral follicle growth in both the follicular and luteal phases of the cycle, which was significantly (P < 0.05) elevated compared with that during menses. For women attending an infertility clinic, those with diagnosed polycystic ovarian syndrome indicated significantly (P < 0.0005) greater circulating INSL3 levels and those with low ovarian reserve showed significantly (P < 0.002) decreased INSL3 values. LIMITATIONS, REASONS FOR CAUTION: These were retrospective studies and the results were obtained from natural cycles only, with their inherent variability. WIDER IMPLICATIONS OF THE FINDINGS: We show for the first time that INSL3 in women does vary across the menstrual cycle, and appears to reflect the number of growing antral follicles recruited within both follicular and luteal phases. STUDY FUNDING/COMPETING INTEREST(S): The present retrospective study was largely supported by departmental funds. There were no competing interests.

Lack of association between serum adiponectin levels and the Pro12Ala polymorphism in Asian Indians

AIMS: The aim of the study was to investigate the association of serum adiponectin levels with the Pro12Ala polymorphism of the peroxisome proliferator activated receptor-gamma (PPARG) gene in Asian Indians. METHODS: We selected 400 diabetic subjects, 200 with the Pro12Pro genotype (100 male and 100 female) and 200 with the Pro12Ala genotype (100 male and 100 female) and 400 age- and sex-matched normal glucose tolerance subjects with similar genotype profiles from the Chennai Urban Rural Epidemiology Study. Fasting serum adiponection levels were measured using radioimmunoassay. The Pro12Ala polymorphism was genotyped by PCR-restriction fragment length polymorphism using BstUI. RESULTS: All clinical and biochemical parameters were similar in the subjects with the Pro12Pro and Pro12Ala genotypes. There was no significant difference in serum adiponectin values between subjects with the Pro12Pro and Pro12Ala genotypes (males 5.4 vs. 5.8 microg/ml, P = 0.546; females 6.9 vs. 7.2 microg/ml, P = 0.748). Adiponectin values did not differ among these two genotypes even when categorized based on their diabetes status (normal glucose tolerance Pro12Pro 7.9 vs. Pro12Ala 7.7 microg/ml, P = 0.994; diabetes Pro12Pro 4.7 vs. Pro12Ala 5.4 microg/ml, P = 0.622). CONCLUSION: The Pro12Ala polymorphism of the PPARG gene is not associated with serum adiponectin levels in Asian Indians.

Modelling of urban sensible heat flux at multiple spatial scales : a demonstration using airborne hyperspectral imagery of Shanghai and a temperature–emissivity separation approach

Urbanization related alterations to the surface energy balance impact urban warming (‘heat islands’), the growth of the boundary layer, and many other biophysical processes. Traditionally, in situ heat flux measures have been used to quantify such processes, but these typically represent only a small local-scale area within the heterogeneous urban environment. For this reason, remote sensing approaches are very attractive for elucidating more spatially representative information. Here we use hyperspectral imagery from a new airborne sensor, the Operative Modular Imaging Spectrometer (OMIS), along with a survey map and meteorological data, to derive the land cover information and surface parameters required to map spatial variations in turbulent sensible heat flux (QH). The results from two spatially-explicit flux retrieval methods which use contrasting approaches and, to a large degree, different input data are compared for a central urban area of Shanghai, China: (1) the Local-scale Urban Meteorological Parameterization Scheme (LUMPS) and (2) an Aerodynamic Resistance Method (ARM). Sensible heat fluxes are determined at the full 6 m spatial resolution of the OMIS sensor, and at lower resolutions via pixel aggregation and spatial averaging. At the 6 m spatial resolution, the sensible heat flux of rooftop dominated pixels exceeds that of roads, water and vegetated areas, with values peaking at ∼ 350 W m− 2, whilst the storage heat flux is greatest for road dominated pixels (peaking at around 420 W m− 2). We investigate the use of both OMIS-derived land surface temperatures made using a Temperature–Emissivity Separation (TES) approach, and land surface temperatures estimated from air temperature measures. Sensible heat flux differences from the two approaches over the entire 2 × 2 km study area are less than 30 W m− 2, suggesting that methods employing either strategy maybe practica1 when operated using low spatial resolution (e.g. 1 km) data. Due to the differing methodologies, direct comparisons between results obtained with the LUMPS and ARM methods are most sensibly made at reduced spatial scales. At 30 m spatial resolution, both approaches produce similar results, with the smallest difference being less than 15 W m− 2 in mean QH averaged over the entire study area. This is encouraging given the differing architecture and data requirements of the LUMPS and ARM methods. Furthermore, in terms of mean study QH, the results obtained by averaging the original 6 m spatial resolution LUMPS-derived QH values to 30 and 90 m spatial resolution are within ∼ 5 W m− 2 of those derived from averaging the original surface parameter maps prior to input into LUMPS, suggesting that that use of much lower spatial resolution spaceborne imagery data, for example from Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) is likely to be a practical solution for heat flux determination in urban areas.

The sensitivity of the tropical circulation and Maritime Continent precipitation to climate model resolution

The dependence of the annual mean tropical precipitation on horizontal resolution is investigated in the atmospheric version of the Hadley Centre General Environment Model (HadGEM1). Reducing the grid spacing from about 350 km to 110 km improves the precipitation distribution in most of the tropics. In particular, characteristic dry biases over South and Southeast Asia including the Maritime Continent as well as wet biases over the western tropical oceans are reduced. The annual-mean precipitation bias is reduced by about one third over the Maritime Continent and the neighbouring ocean basins associated with it via the Walker circulation. Sensitivity experiments show that much of the improvement with resolution in the Maritime Continent region is due to the specification of better resolved surface boundary conditions (land fraction, soil and vegetation parameters) at the higher resolution. It is shown that in particular the formulation of the coastal tiling scheme may cause resolution sensitivity of the mean simulated climate. The improvement in the tropical mean precipitation in this region is not primarily associated with the better representation of orography at the higher resolution, nor with changes in the eddy transport of moisture. Sizeable sensitivity to changes in the surface fields may be one of the reasons for the large variation of the mean tropical precipitation distribution seen across climate models.

Forecasting with difference and trend stationary models

Although difference-stationary (DS) and trend-stationary (TS) processes have been subject to considerable analysis, there are no direct comparisons for each being the data-generation process (DGP). We examine incorrect choice between these models for forecasting for both known and estimated parameters. Three sets of Monte Carlo simulations illustrate the analysis, to evaluate the biases in conventional standard errors when each model is mis-specified, compute the relative mean-square forecast errors of the two models for both DGPs, and investigate autocorrelated errors, so both models can better approximate the converse DGP. The outcomes are surprisingly different from established results.