Root Cause Analysis by a Combined Sparse Classification and Monte Carlo Approach

Classification methods with embedded feature selection capability are very appealing for the analysis of complex processes since they allow the analysis of root causes even when the number of input variables is high. In this work, we investigate the performance of three techniques for classification within a Monte Carlo strategy with the aim of root cause analysis. We consider the naive bayes classifier and the logistic regression model with two different implementations for controlling model complexity, namely, a LASSO-like implementation with a L1 norm regularization and a fully Bayesian implementation of the logistic model, the so called relevance vector machine. Several challenges can arise when estimating such models mainly linked to the characteristics of the data: a large number of input variables, high correlation among subsets of variables, the situation where the number of variables is higher than the number of available data points and the case of unbalanced datasets. Using an ecological and a semiconductor manufacturing dataset, we show advantages and drawbacks of each method, highlighting the superior performance in term of classification accuracy for the relevance vector machine with respect to the other classifiers. Moreover, we show how the combination of the proposed techniques and the Monte Carlo approach can be used to get more robust insights into the problem under analysis when faced with challenging modelling conditions.

BUCKLED LAYER STRUCTURE FOR ATOMIC ADSORPTION ON W(100) - THE (ROOT-2X-ROOT-2)R45-DEGREES NITROGEN STRUCTURE FROM ATLEED

Adsorption of 0.5 monolayer of N adatoms on W{100} results in a sharp (root 2 X root 2)R45 degrees LEED pattern. The only previous quantitative LEED study of this system gave a simple overlayer model with a Pendry R-factor of 0.55. An exhaustive search has been made of possible structures, including a novel vacancy reconstruction, displacive reconstructions and underlayer adsorption. From this work a new overlayer structure is derived with an R(p) value of 0.22, displaying a considerable buckling of 0.27 +/- 0.05 Angstrom within the second W layer and consequently involving large changes in the interlayer spacings of the surface. The N adatom is pseudo-five-fold coordinated to the W surface, bonding to a second-layer W atom with a nearest-neighbour bond length of 2.13 Angstrom and with the four next-nearest-neighbour W atoms in the surface plane at 2.27 Angstrom. The structure does not resolve the work function anomaly observed on this surface.

Calluna vulgaris root cells show increased capacity for amino acid uptake when colonized with the mycorrhizal fungus Hymenoscyphus ericae

Ericoid mycorrhizas are believed to improve N nutrition of many ericaceous plant species that typically occur in habitats with impoverished nutrient status, by releasing amino acids from organic N forms. Despite the ubiquity of mycorrhizal formation the mechanisms and regulation of nutrient transport in mycorrhizal associations are poorly understood. We used an electrophysiological approach to study how amino acid transport characteristics of Calluna vulgaris were affected by colonization with the ericoid mycorrhiza fungus Hymenoscyphus ericae. Both the V_max and K_m parameters of amino acid uptake were affected by fungal colonization in a manner consistent with an increased availability of amino acid to the plant. The ecophysiological significance of altered amino acid transport in colonized root cells of C. vulgaris is discussed. © New Phytologist (2002).

Genetic diversity of root-associated fungal endophytes from Calluna vulgaris at contrasting field sites

A total of 107 putative ericoid mycorrhizal endophytes were isolated from hair roots of Calluna vulgaris from two abandoned arsenic/copper mine sites and a natural heathland site in southwest England. The endophytes were initially grouped as 14 RFLP types, based on the results of ITS-RFLP analysis using the restriction endonucleases Hinf I, Rsa I and Hae III. ITS sequences were obtained for representative isolates from each RFLP type and compared phylogenetically with sequences for known ericoid mycorrhizal endophytes and selected ascomycetes. The majority of endophyte isolates (62-92%) from each site were identified as Hymenoscyphus ericae, but a number of other less common mycorrhizal RFLP types were also identified, all of which appear to have strong affinities with the order Leotiales. None of the less common RFLP types was isolated from C. vulgaris at more than one field site. Neighbour-joining analysis indicated similarities between the endophytes from C. vulgaris and mycorrhizal endophytes isolated from other Ericaceae and Epacridaceae hosts in North America and Australia.

Cable correction of membrane currents recorded from root hairs of Arabidopsis thaliana L.

Membrane currents were recorded under voltage clamp from root hairs of Arabidopsis thaliana L. using the two-electrode method. Concurrent measurements of membrane voltage distal to the point of current injection were also carried out to assess the extent of current dissipation along the root hair axis. Estimates of the characteristic cable length, λ, showed this parameter to be a function both of membrane voltage and of substrate concentration for transport. The mean value for λ at 0 mV was 103 ± 20 μm (n=17), but ranged by as much as 6-fold in any one cell for membrane voltages from -300 to +40 mV and was affected by 0.25 to 3-fold at any one voltage on raising [K⁺]₀ from 0.1 to 10 mol m^-3. Current dissipation along the length of the cells lead to serious distortions of the current-voltage [I-V) characteristic, including consistent underestimates of membrane current as well as a general linearization of the I-V curve and a masking of conductance changes in the presence of transported substrates. In some experiments, microelectrodes were also placed in neighbouring epidermal cells to record the extent of intercellular coupling. Even with current-passing microelectrodes placed at the base of root hairs, coupling was ≤5% (voltage deflection of the epidermal cell ≤5% that recorded at the site of current injection), indicating an appreciable resistance to current passage between cells. These results demonstrate the feasibility of using root hairs as a 'single-cell model' in electrophysiological analyses of transport across the higher-plant plasma membrane; they also confirmed the need to correct for the cable properties of these cells on a cell-by-cell basis. © 1994 Oxford University Press.

A novel method of quantifying root exudation in the presence of soil microflora

A microcosm is described in which root exudation may be estimated in the presence of microorganisms. Ryegrass seedlings are grown in microcosms in which roots were spatially separated from a microbial inoculant by a Millipore membrane. Seedlings grown in the microcosms were labelled with [¹⁴C]-CO₂, and the fate of the label within the plant and rhizosphere was determined. Inoculation of the microcosms with Cladosporium resinae increased net fixation of the [¹⁴C] label compared to plants grown under sterile conditions. Inoculation also increased root exudation. The use of the microcosm was illustrated and its applications discussed. © 1991 Kluwer Academic Publishers.

Root caries development in older adults- a prospective cohort study

Objectives: To identify factors associated with root caries development during a two year period in a population of independently living older adults. Methods: A prospective cohort study was carried out with 334 independently living volunteers aged 65 and older. At baseline (t0), each participant completed a questionnaire which recorded age, gender, medical history, fluoride exposure, oral and denture hygiene practices, smoking and alcohol consumption, diet information, and socio economic information. Clinical examinations were performed and stimulated saliva samples were collected. Patients were reviewed 12(t1) and 24(t2) months later to determine the root caries increment. Results: 307 adults were assessed at t1 and 280 were assessed at t2 with 83.8% of participants examined at 24 months. Incidence of root caries in this cohort was 17.4% at t1 and 21.6% at t2. The mean root caries increment was 0.43 (SD 1.45) surfaces at t1 and 0.70 (SD 1.86) surfaces at t2. Age >70 years, completing education at primary level, poor oral hygiene, xerostomia, coronal decay at baseline, higher root caries index at baseline and number of exposed root surfaces showed a statistically significant association (P < 0.05) with root caries development. Conclusion: Root caries is a substantive dental health problem for our older population. Root caries prevention strategies should be targeted at older adults who have poor plaque control and high levels of caries experience. In particular patients with xerostomia should be targeted with preventive measures.

Risk indicators associated with root decay in independently living older adults

Objective: To determine the risk indicators associated with root caries experience in a cohort of independently living older adults in Ireland.
Methods: The data reported in the present study were obtained from a prospective longitudinal study conducted on the risk factors associated with root caries incidence in a cohort of independently living older adults (n=334). Each subject underwent an oral examination, performed by a single calibrated examiner, to determine the root caries index and other clinical variables. Questionnaires were used to collect data on oral hygiene habits, diet, smoking and alcohol habits and education level. A regression analysis with the outcome variable of root caries experience (no/yes) was conducted.
Results: A total of 334 older adults with a mean age of 69.1 years were examined. 53.3% had at least one filled or decayed root surface. The median root caries index was 3.13 (IQR 0.00, 13.92). The results from the multivariate regression analysis indicated that individuals with poor plaque control (OR 9.59, 95%CI 3.84-24.00), xerostomia (OR 18.49, 95%CI 2.00-172.80), two or more teeth with coronal decay (OR 4.50, 95% CI 2.02-10.02) and 37 or more exposed root surfaces (OR 5.48, 95% CI 2.49-12.01) were more likely to have been affected by root caries.
Conclusions: The prevalence of root caries was high in this cohort. This study suggests a correlation between root caries and the variables poor plaque control, xerostomia, coronal decay (≥2 teeth affected) and exposed root surfaces (≥37). The significance of these risk indicators and the resulting prediction model should be further evaluated in a prospective study of root caries incidence.
Clinical Significance: Identification of risk indicators for root caries in independently living older adults would facilitate dental practitioners to identify those who would benefit most from interventions aimed at prevention.