Relative Importance Of Variables Research Articles

Yield and Properties Prediction Based on the Multicondition LSTM Model for the Solvent Deasphalting Process.

Solvent deasphalting (SDA) is a complex multiscale continuous process. The operation mode of the SDA process is not considered in the related data-driven model. Therefore, this paper proposes a time lag process prediction model with multiple operation modes to solve the above problem. First, based on random forests, the relative importance of initial input variables in the SDA process on DAO yield and Conradson carbon residual are studied and features are selected according to the results. Then, the stack denoising autoencoder (SDAE) is used to reconstruct the data and obtain the nonlinear mapping information of hidden layers of SDAE and achieve feature dimension reduction. SDAE can improve clustering accuracy of fuzzy c-means, and the operation mode of SDA process is accurately divided. Long short-term memory (LSTM) is used to establish a multicondition LSTM model. Compared with the traditional LSTM model, the multicondition LSTM model has a higher prediction accuracy with R 2 > 0.95. The sensitivity analyses of the properties of feed and operating conditions on DAO yield are consistent with the principle of two-phase countercurrent extraction in the SDA process. In addition, the benchmark test of the Tennessee Eastman process shows that the proposed method is also effective in the fault detection of other processes. Because the multicondition LSTM can predict the future process measurement data according to operating mode, it can better avoid the false alarm problem and predict the fault earlier.

Examining the use of decision trees in population health surveillance research: an application to youth mental health survey data in the COMPASS study.

In population health surveillance research, survey data are commonly analyzed using regression methods; however, these methods have limited ability to examine complex relationships. In contrast, decision tree models are ideally suited for segmenting populations and examining complex interactions among factors, and their use within health research is growing. This article provides a methodological overview of decision trees and their application to youth mental health survey data. The performance of two popular decision tree techniques, the classification and regression tree (CART) and conditional inference tree (CTREE) techniques, is compared to traditional linear and logistic regression models through an application to youth mental health outcomes in the COMPASS study. Data were collected from 74 501 students across 136 schools in Canada. Anxiety, depression and psychosocial well-being outcomes were measured along with 23 sociodemographic and health behaviour predictors. Model performance was assessed using measures of prediction accuracy, parsimony and relative variable importance. Decision tree and regression models consistently identified the same sets of most important predictors for each outcome, indicating a general level of agreement between methods. Tree models had lower prediction accuracy but were more parsimonious and placed greater relative importance on key differentiating factors. Decision trees provide a means of identifying high-risk subgroups to whom prevention and intervention efforts can be targeted, making them a useful tool to address research questions that cannot be answered by traditional regression methods.

Abstract WP94: Prediction Of Stroke Thrombectomy Outcomes From The Neurovascular Quality Initiative-quality Outcomes Database (NVQI-QOD) Registry Using Probabilistic Learning Models

Introduction: Mechanical Thrombectomy (MT) is standard-of-care interventional management of Acute Ischemic Stroke (AIS) due to large vessel occlusions. However, patient outcomes remain variable after intervention, with unclear optimization strategies for patient selection. The NVQI-QOD registry documents detailed patient characteristics, pre-operative imaging, procedure metrics, and post-operative outcomes of neurointerventional procedures. Data are highly informative, but there is inherent uncertainty in all medical interventions. We introduce a probabilistic learning model that predicts the expected distribution of MT outcomes. Methods: We identified two groups of variables from the NVQI-QOD AIS Thrombectomy registry: 1) data available at the time of MT (Group-preop), and 2) data available 24 h after MT, which can be valuable for prognostication (Group-24h). After filtering missing values, there were 1174 and 1405 examples in the two groups, respectively. In each group, training and test datasets were split using a ratio of 8:2. A probabilistic Neural Network (NN) ( Fig. 1a ) was developed to predict the distribution of changes in pre- and post-MT NIH Stroke Scale (NIHSS) ( y ), and it was trained using both groups of variables as inputs. Results: After training, the probabilistic network accurately described the distributions of changes in the NIHSS (represented by predicted means and SDs) based on input variables ( Fig. 1b - upper ). Notably, in Group-preop, even patients with the worst predicted outcomes had an approximately 50% chance of improvement. Fig. 1b - lower demonstrates the relative importance of variables to the NN. Conclusions: This NN model demonstrates the utility of probabilistic learning in clinical decision-making and prognosis. Our results reinforce the substantial benefits of MT, that can still improve outcomes in nearly half of patients with the worst predicted change in NIHSS on pre-operative analysis.

XAIRE: An ensemble-based methodology for determining the relative importance of variables in regression tasks. Application to a hospital emergency department

Nowadays it is increasingly important in many applications to understand how different factors influence a variable of interest in a predictive modeling process. This task becomes particularly important in the context of Explainable Artificial Intelligence. Knowing the relative impact of each variable on the output allows us to acquire more information about the problem and about the output provided by a model.This paper proposes a new methodology, XAIRE, that determines the relative importance of input variables in a prediction environment, considering multiple prediction models in order to increase generality and avoid bias inherent in a particular learning algorithm. Concretely, we present an ensemble-based methodology that promotes the aggregation of results from several prediction methods to obtain a relative importance ranking. Also, statistical tests are considered in the methodology in order to reveal significant differences between the relative importance of the predictor variables. As a case study, XAIRE is applied to the arrival of patients in a Hospital Emergency Department, which has resulted in one of the largest sets of different predictor variables in the literature. Results show the extracted knowledge related to the relative importance of the predictors involved in the case study.

Changes in ImPACT Cognitive Subtest Networks Following Sport-Related Concussion.

High school athletes are administered ImPACT at the start of the academic year or sport season and again after suspected concussion. Concussion management involves the comparison of baseline and post-injury cognitive scores with declines in scores providing evidence for concussive injury. A network framework may provide additional information about post-concussive cognitive changes and expand characterization of sport-related concussion (SRC) recovery. Retrospective cohort study. High school. High school athletes (n = 1553) were administered ImPACT at baseline (T1), post-SRC (T2 = 72 h of injury), and prior to return to play (T3 = within two weeks post-injury). ImPACT cognitive subtest scores. Cognitive networks were calculated and compared over three time points. Centrality indices were calculated to determine the relative importance of cognitive variables within networks. Network connectivity increased from T1 to T2 and remained hyperconnected at T3. There was evidence of network reorganization between T1 and T3. Processing speed was central within each network, and visual memory and impulsivity became more central over time. The results suggest potential evidence of cognitive network change over time. Centrality findings suggest research specific to visual memory and impulse control difficulties during the post-concussion recovery period is warranted. Network analysis may provide additional information about cognitive recovery following SRC and could potentially serve as an effective means of monitoring persisting cognitive symptoms after concussion.

Mapping the strength of agro-ecological lightweight concrete containing oil palm by-product using artificial intelligence techniques

The critical challenge for the cement production industry is the high emission of greenhouse gases. For the sustainability in a cycling economy context civil and environmental engineers should reconsider future generations and should redesign and manufacturing in a safe and eco-friendly alternative. Previous experimental studies have shown, that using oil palm by-product in the production of sustainable lightweight structural concrete, (SLSC) presents good life cycle, land-use efficiency and economic benefits. In this work a formula-based model is proposed to evaluate the strength of SLSC, knowledge-aided modeling based on machine learning methods (e. g., for further development, Multivariate Adaptive Regression Splines (MARS), Gene Expression Programming (GEP), M5P Model Tree (M5P) and Extreme Learning Machine (ELM)) were selected. A comprehensive and reliable experimental record comprising of 449 data of SLSC containing oil palm by-product was utilized to formulate strength predictive models. The proposed models were investigated, in terms of assessment criteria, error measurements, and performance evaluation. The efficiency of the MARS model was shown by comparing its performance with GEP, M5P and ELM. As well as Monte-Carlo simulation approach was conducted to verify the proposed models. Evaluation of the relative importance of effective variables presented that the gravel (11.84%), oil palm waste aggregate (11.38%) and water to binder ratio (11.15%) were the most influential features in the strength prediction of the SLSC. Also, a parametric study is developed to provide robustness of the proposed predictive model. Results of the modeling provide a new understanding of the provision of an auto-estimated strength evaluation of SLSC incorporating oil palm by-product with superiority promotion of knowledge-aided modeling mechanism.

Estimating stay cable vibration under typhoon with an explainable ensemble learning model

Excessive vibration of stay cables in strong winds has been a concern for bridge operators, which impairs the durability of both the cables and the bridge structure. This paper develops a data-driven approach to predict the amplitude of the cable vibration using an ensemble learning model. The model aims to predict cable vibrations in both in-plane and out-of-plane directions, with the wind speed, wind direction, turbulence intensity, and deck acceleration as input variables. Especially, the deck acceleration is included considering the deck-cable interaction and vehicle effects, which significantly improved the accuracy of the prediction. Furthermore, the model is interpreted with local interpretable model-agnostic explanations (LIME) and partial dependence plot (PDP) methods. The former demonstrates the relative importance of input variables on a global scale, and the latter indicates the correlation between individual input variables with the prediction target. The investigation is validated using the data harnessed from structural health monitoring (SHM) of a 1088-m cable-stayed bridge during three typhoon events. The adopted Gradient boosting regression tree (GBRT) model demonstrated better performance than other state-of-the-art machine learning models. The developed approach can provide guidance on preventive maintenance of stay cables to avoid damage due to excessive vibration.

Most patients with non-hypertensive diseases at a critical care resuscitation unit require arterial pressure monitoring: a prospective observational study.

Blood pressure (BP) monitoring is essential for patient care. Invasive arterial BP (IABP) is more accurate than non-invasive BP (NIBP), although the clinical significance of this difference is unknown. We hypothesized that IABP would result in a change of management (COM) among patients with non-hypertensive diseases in the acute phase of resuscitation. This prospective study included adults admitted to the Critical Care Resuscitation Unit (CCRU) with non-hypertensive disease from February 1, 2019, to May 31, 2021. Management plans to maintain a mean arterial pressure >65 mmHg (1 mmHg=0.133 kPa) were recorded in real time for both NIBP and IABP measurements. A COM was defined as a discrepancy between IABP and NIBP that resulted in an increase/decrease or addition/discontinuation of a medication/infusion. Classification and regression tree analysis identified significant variables associated with a COM and assigned relative variable importance (RVI) values. Among the 206 patients analyzed, a COM occurred in 94 (45.6% [94/206]) patients. The most common COM was an increase in current infusion dosages (40 patients, 19.4%). Patients receiving norepinephrine at arterial cannulation were more likely to have a COM compared with those without (45 [47.9%] vs. 32 [28.6%], P=0.004). Receiving norepinephrine (relative variable importance [RVI] 100%) was the most significant factor associated with a COM. No complications were identified with IABP use. A COM occurred in 94 (45.6%) non-hypertensive patients in the CCRU. Receiving vasopressors was the greatest factor associated with COM. Clinicians should consider IABP monitoring more often in non-hypertensive patients requiring norepinephrine in the acute resuscitation phase. Further studies are necessary to confirm the risk-to-benefit ratios of IABP among these high-risk patients.

Classification and prediction of the effects of nutritional intake on diabetes mellitus using artificial neural network sensitivity analysis: 7th Korea National Health and Nutrition Examination Survey.

This study aimed to predict the association between nutritional intake and diabetes mellitus (DM) by developing an artificial neural network (ANN) model for older adults. Participants aged over 65 years from the 7th (2016-2018) Korea National Health and Nutrition Examination Survey were included. The diagnostic criteria of DM were set as output variables, while various nutritional intakes were set as input variables. An ANN model comprising one input layer with 16 nodes, one hidden layer with 12 nodes, and one output layer with one node was implemented in the MATLAB® programming language. A sensitivity analysis was conducted to determine the relative importance of the input variables in predicting the output. Our DM-predicting neural network model exhibited relatively high accuracy (81.3%) with 11 nutrient inputs, namely, thiamin, carbohydrates, potassium, energy, cholesterol, sugar, vitamin A, riboflavin, protein, vitamin C, and fat. In this study, the neural network sensitivity analysis method based on nutrient intake demonstrated a relatively accurate classification and prediction of DM in the older population.

Optimum Weather Conditions for Prescribed Burning in the Flint Hills Region

Highlights The subjectivity of landowner decisions in prescribed burning was influenced by solar radiation and relative humidity. Landowners preferred to burn when there was less cloud cover, which promoted O3 generation. Proposed optimum weather conditions will reduce the smoke impact on O3. Abstract. Land managers use weather conditions to determine when meteorological and fuel conditions are suitable for prescribed burning. The objectives of this study were: 1) to characterize meteorological parameters’ sensitivities and identify the influential weather parameters connected to daily fire activities; and 2) to identify optimum weather conditions that are practical for mitigating the smoke impact associated with prescribed fires in the Flint Hills region. Machine learning techniques using random forest (RF) with Shapley additive explanation (SHAP) values and Pearson correlation were used to identify the relative importance of weather variables and their impact on landowner burn decisions in the Flint Hills region based on historic data from 2003-2019. Results showed that less cloud cover demonstrated superior predictive power in landowner decisions for prescribed burning and largely resulted in elevated ambient ozone (O3). We found out that six days per burning season on average provided the proposed optimum weather conditions, where cloud cover ranged from 10+ to 55%. However, 62% of the total burned area occurred when the cloud cover was less than 10+% during 2003-2019. By changing the current burning occurrence to the proposed optimum weather conditions in our study, the probability of 70+ ppb O3 occurrence could be reduced from 35% to 9% on heavy fire days. Keywords: Cloud cover, Optimum weather conditions, Ozone, Prescribed burning.

Predicting the Thickness of an Excavation Damaged Zone around the Roadway Using the DA-RF Hybrid Model

After the excavation of the roadway, the original stress balance is destroyed, resulting in the redistribution of stress and the formation of an excavation damaged zone (EDZ) around the roadway. The thickness of EDZ is the key basis for roadway stability discrimination and support structure design, and it is of great engineering significance to accurately predict the thickness of EDZ. Considering the advantages of machine learning (ML) in dealing with high-dimensional, nonlinear problems, a hybrid prediction model based on the random forest (RF) algorithm is developed in this paper. The model used the dragonfly algorithm (DA) to optimize two hyperparameters in RF, namely mtry and ntree, and used mean absolute error (MAE), root mean square error (RMSE), determination coefficient (R2), and variance accounted for (VAF) to evaluate model prediction performance. A database containing 217 sets of data was collected, with embedding depth (ED), drift span (DS), surrounding rock mass strength (RMS), joint index (JI) as input variables, and the excavation damaged zone thickness (EDZT) as output variable. In addition, four classic models, back propagation neural network (BPNN), extreme learning machine (ELM), radial basis function network (RBF), and RF were compared with the DA-RF model. The results showed that the DARF mold had the best prediction performance (training set: MAE = 0.1036, RMSE = 0.1514, R2 = 0.9577, VAF = 94.2645; test set: MAE = 0.1115, RMSE = 0.1417, R2 = 0.9423, VAF = 94.0836). The results of the sensitivity analysis showed that the relative importance of each input variable was DS, ED, RMS, and JI from low to high.

Structuring forces of intertidal infaunal communities on the northern coast of British Columbia, Canada: assessing the relative importance of top-down, bottom-up, middle-out, and abiotic variables

While trophic and habitat-related abiotic variables (predation, competition, tolerance, etc.) are known to influence community structure in many ecosystems, some systems appear to be only minimally influenced by these variables. Sampling multiple tidal flat communities in northern BC, Canada, we investigated the relative importance of top-down and middle-out (mesopredators) variables, competition for resources (bottom up), and abiotic variables in structuring an infaunal community (invertebrates living in sediment). Similar to previous studies on mudflats in the Bay of Fundy (also at a north temperate latitude), we determined that these variables accounted for a minor (0%–9%) proportion of the observed variation in this infaunal community, suggesting that these variables play a small role in structuring this community. Based on the results of our study and in combination with previous experiments on infaunal recovery patterns post disturbance, we posit that the main factors influencing these infaunal communities likely operate at a scale of sites (kilometres) and(or) plot (metres or less) but not transects (10–100 m within site). Candidate forces structuring these intertidal communities that need future examination include regional species pools and the variables that affect these pools, sediment biogeochemistry, and disturbance/recovery history of a site. The similarity of our Pacific coast findings to those from the north temperate Atlantic coast suggests some similarity in the processes structuring these distinct infaunal communities.

Analysis of the evolution of a fisheries management plan based on environmental governance: Living laboratory in the Olo River, Portugal

Environmental governance aims to develop and understand the lessons, achievements and challenges of planning at both local and global levels. In particular, integrated river management is key to the sustainable development of ecosystems, and the collaborative and inclusive approach among stakeholders supports the decision-making process. The objective of this study was to assess the current management model for recreational fisheries in the Olo River Reserved Fishing Zone (RFZ). This management model aims to better rationalise the use of fishery resources by allowing only recreational fishing, ensuring the sustainability of this activity and of the environment, and guaranteeing that this activity contributes to the development of the basin. To this end, the state of this activity was evaluated before and after the implementation of the First Management Plan and the Second Management Plan. Samples of ichthyofauna were taken and an analysis of the data on the age, growth and physical condition of the endemic species was carried out. Multivariate analysis techniques applied to the biological and physical data sets allowed the determination of distribution patterns of Olo Basin and the relative importance of environmental variables in the discrimination of the fish assemblages.The presence of the species Salmo trutta fario improved in the stations sampled throughout the study. The abundance of fish throughout the study showed a remarkable improvement, from 1229 (year 2008), 1142 (year 2013) and 1781 (year 2018), the drop in 2013 was being due to a wildfire in the area. This fishing area has been a true living laboratory for the development of new and more sustainable inland fisheries management practices. The development of this planning and management model was successfully exported to other river basins in the area.

Husband-wife Roles as a Correlate of Contraceptive and Fertility Behaviour

In this paper, an investigation of reproductive behaviour within the socioeconomic and cultural frameworks is carried out to find the extent to which socioeconomic, cultural, and attitudinal variables (such as husband and wife’s education, family income, husband’s occupation, child mortality, exposure to the mass media, and husbandand- wife relationship in terms of egalitarian roles, role-segregation, husband’s authority, and domination in family and non-family decisions) influence the fertility decision-making process. The quantitative and qualitative techniques are used for exploring the respondents’ views regarding contraceptive and fertility behaviour. Principle Component Analysis (PCA) is applied to identify new meaningful underlying variables and to reduce the multi-dimensionality of variables. The chi-square test is employed to explore the relationships between the predictor variables and the dependent variables. Multiple linear regression is also used to establish the relative importance of each of the predictor variables. Bivariate, multiple linear regression and qualitative analysis demonstrate that preferences for smaller families and contraceptive use were found to be consistently associated with modern attitudes and behaviour towards the husband-and-wife relationship. Family income, husband’s occupation, child mortality, and age at marriage offered no explanation of the reproductive behaviour. It is concluded that cultural setting and tradition exert an important influence on reproductive behaviour independent of development in economic realities. It is suggested that for the attainment of demographicdevelopmental objectives, the issue of women’s status is not incidental; it is essential. The argument is not that improvements in women’s status need to be pursued only for population policy purposes, but rather that they comprise a crucial social developmental goal in their own right.

Design of experiments optimization of fluorine-doped tin oxide films prepared by spray pyrolysis for photovoltaic applications

Transparent conducting oxides (TCOs) have witnessed an ever-expanding use in our lives through many optoelectronic applications, namely photovoltaic (PV) devices. Indium-tin oxide (ITO) is the most used and studied TCO, but it lacks thermal and chemical stability and indium is a scarce and toxic element. Fluorine-doped tin oxide (FTO) emerged as the most promising alternative to ITO, presenting better thermal and chemical stability. Among the numerous techniques for depositing FTO thin films over glass substrates, spray pyrolysis is the simplest and most economical, with great potential for upscaling. However, the relative importance of the experimental variables that influence the optoelectronic properties remains barely addressed. Following this premise, the present work aimed at optimizing the deposition of FTO films on soda lime glass (SLG) substrates by spray pyrolysis following a Design of Experiments (DoE) methodology. The optoelectronic properties of FTO-SLG substrates was evaluated based on their optical transmittance and sheet resistance, both combined in a figure of merit (FoM) tailored for PV applications. It was concluded that the volume of sprayed solution and the fluorine/tin ratio in the precursor have the greatest influence in the FoM, being the optimal deposition conditions a sprayed volume of 60.8 ml and a [F]/[Sn] ratio of 0.45. FTO-SLG substrates prepared with these conditions achieved a FoM of 0.680 Ω□−1/10, corresponding to a sheet resistance of 3.40 Ω□ and a transmittance equivalent to 77% of the maximum current generated in the considered spectrum. The improved FoM was validated in dye-sensitized and perovskite solar cells (DSSCs and PSCs, respectively), assembled with in-house optimized and commercial substrates. The efficiency of DSSCs was improved by 8.9% (relative), whereas PSCs achieved a light-to-power efficiency of 17% (absolute), corresponding to an improvement of 4.7% (relative).

OLDER ADULT PREFERENCES FOR THE FORMAT AND CONTENT OF PRESCRIPTION MEDICATION LABELS

Abstract The risk of misinterpreting prescription medicine labels (PMLs) is higher among older adults. Given the limited space on PMLs, their format and content should cater to older adults’ medication information and legibility needs. We established the ranking of seven medication-related content variables (indication, precautions, interaction or paired medicines, food instructions, side effects, expiry date, and missed dose action; provision of these content is currently not legally mandated) to be prioritised on PMLs using best-worst scaling (BWS), and determined the relative utilities of five format attributes (presentation of dose-frequency instructions, font size, presentation of dose, presentation and colour of precautions) when incorporated onto PMLs using discrete choice experiment (DCE), from the perspective of older Singaporeans. The variables and attributes were informed by our prior qualitative study and best practices guidelines for PMLs. We recruited 280 older adults (mean age: 68.8±5.4 years). A BWS object case was used to evaluate the relative importance of medication-related content variables (range: 0 to 1). A mixed logit model estimated the relative utilities of attribute levels, which enabled calculation of importance scores for ranking the format attributes (range: 0% to 100%). The three most-preferred content variables were indication (1.000), precautions (0.853) and interaction or paired medicines (0.845). The top three format preferences were tabular presentation of dose-frequency instructions (43.7%), large font size (22.3%) and precautions in red colour (13.6%). The findings inform an actionable way to improve the content and format of PMLs in the real-world, which is aligned with the stated priorities of older adults.

Spatiotemporal variability and controlling factors of indirect N2O emission in a typical complex watershed

The mechanisms of N2O emissions from inland rivers remain poorly understood because of the high variability of dissolved N2O concentration and the complexity of influencing factors. Thus, it is necessary to research the spatiotemporal patterns and influencing factors to understand the driving mechanisms of riverine N2O emissions. We combine the Soil and Water Assessment Tool (SWAT) outputs with established empirical equations to identify the spatiotemporal fluctuations of riverine N2O emissions and evaluate model performance through field measurements in a typical watershed. The spatiotemporal hotspots of N2O emissions are then determined, and the relative importance of environmental variables is further determined by correlation and attribution analysis. The results indicate that the riverine N2O emissions are relatively high from August to October, which accounted for 35.38% of the annual emissions. Temporal changes are attributed to agricultural activities and meteorological factors. Agricultural activities such as planting and fertilization lead to increased diffuse nitrogen loads on the land surface. Meantime, heavy precipitation events enhance the transport of nutrients, resulting in changes in nitrogen levels in the river. Spatial analysis shows that the urban watersheds (191.22 ± 156.19 μmol m−2 d−1) are the hotspots of riverine N2O emission, which are 1.55–3.03 times that of non-urban rivers. Spatial variations are mainly affected by riverine physicochemical indicators for different watersheds. Sewage from various sources received by urban rivers provides appropriate environmental conditions for N2O production, and transports large exogenous dissolved N2O. Furthermore, salinity (r = 0.80; p < 0.001) and nitrogen nutrients in riverine physicochemical indicators show a significant correlation with N2O fluxes. It emphasizes that N-related (TN, NH4+, NO3−) indicators are important reactants for N2O generation, which can promote nitrification and denitrification. Meanwhile, the results of structural equation modeling (SEM) also demonstrate that N2O emissions follow a similar pattern to riverine dissolved N2O concentration (r = 0.841, p < 0.001), and non-point source (r = 0.678, p < 0.001) play an important role in the changes of dissolved N2O concentrations. Our results highlight that certain hot moments and hot spots of rivers play a disproportionate role in year-round and basin-wide N2O emissions, respectively. It is necessary to implement more effective management measures by controlling key environmental factors to reduce N2O emissions.

Optimization of Grinding Parameters of Tool Steel by the Soft Computing Technique.

Grinding is one of the most complex and accurate machining processes, and the efficiency of the grinding wheel depends significantly on its surface properties. This work aims to propose an algorithmic manner that reduces the cost and time to conduct grinding of an optimized DIN 1.2080 tool steel (SPK) using a soft computing technique to obtain the best combination of input parameters including depth of cut (20, 40, 60 μm), wheel speed (15, 20, 25 m/s), feed rate (100, 300, 500 mm/s), and incidence angle (0, 30, 45 de grees) with respect to output parameters consisting of average surface roughness and specific grinding energy. According to the input parameters and their levels, an experiment using fractional factorial design of experiment (RFDOE) was designed. Later on, two parallel feed-forward backpropagation (FFBPNN) networks with similar topology made up of 4, 11, and 1 units in their input, hidden, and output layers are trained, respectively. After sensitivity analyses of networks for determination of the relative importance of input variables, a genetic algorithm (GA) adopting linear programming (LP) based on Euclidean distance is coupled to networks to seek out the best combinations of input parameters that result in minimum average surface roughness and minimum specific grinding energy. The findings revealed that RFDOE provides valid data for training FFBP networks with a total goodness value of more than 1.99 in both cases. The sensitivity analyses showed that feed rate (38.97%) and incidence angle (33.94%) contribute the most in the case of average surface roughness and specific grinding energy networks, respectively. Despite the similar surface quality based on scanning electron microscopy (SEM), the optimization resulted in an optimized condition of the depth of cut of 25.23 μm, wheel speed of 15.02 mm/s, feed rate of 369.45 mm/s, and incidence angle of 44.98 de grees, which had a lower cost value (0.0146) than the optimum one (0.0953). Thus, this study highlights that RFDOE with a hybrid optimization using FFBP networks-GA/LP can effectively minimize both average surface roughness and specific grinding energy of SPK.

A Random Forest Classification Model for Spasticity after Stroke

<h3>Research Objectives</h3> The objective of this analysis was to developed a Random Forest Classification Model to determine which sociodemographic and clinical health characteristics predict the development of post-stroke spasticity (PSS). <h3>Design</h3> Random Forest Classification Modelling with data collected from a retrospective review of patients' medical charts and the National Rehabilitation Reporting System. <h3>Setting</h3> Specialized stroke rehabilitation inpatient and outpatient units from a large, urban centre in Ontario, Canada. <h3>Participants</h3> A total of 922 individuals (102 spastic) attended stroke rehabilitation between July 1, 2015 and June 30, 2020. A synthetic minority oversampling technique was used to generate a dataset of 1632 individuals (408 spastic). <h3>Interventions</h3> None. <h3>Main Outcome Measures</h3> Clinical health data (i.e., number, type and location of stroke, age at stroke), length of stay, LOS); presence of spasticity, presence of aphasia, endovascular or tissue plasminogen activator treatment, Chedoke-McMaster impairment inventory scores, Functional Independence Measure motor and cognitive sub-scores. <h3>Results</h3> Based on the results of a feature hyper-parameterization procedure, a model with 1501 trees was selected with 8 variables (mtry=8). A random forest model was then trained using these parameters on 80% of the dataset. The relative importance of predictor variables was assessed by examining the decrease in model accuracy when a particular variable is permutated. Chedoke McMaster impairment scores, age at stroke onset, Functional Independence Measure motor sub-scores and length of stay appear to be the most important predictor variables when compared to the remaining 7. When testing the model performance on the untrained 20% of the dataset, the model performed with a balanced accuracy of 82.5%, a Cohen's kappa value of 0.71 and a spasticity classification error of 32.1%. <h3>Conclusions</h3> Overall, our model performed relatively well, but could be tuned more for sensitivity to prioritize the correct identification of spasticity. In the future, more data will be required to validate the generalizability of the model and robustness of fit, as well as to eliminate the need for synthetically created data. <h3>Author(s) Disclosures</h3> None.

Elevation drives taxonomic, functional and phylogenetic β‐diversity of phyllostomid bats in the Amazon biome

AbstractAimWe evaluated the relative importance of geographical and environment variables for taxonomic, phylogenetic and functional β‐diversity of phyllostomid bats along the entire Amazon biome and specifically in the lowlands.LocationAmazon biome.TaxonChiroptera.MethodsWe carried out a bibliographic review and compiled a wide and unprecedented database of 106 phyllostomid bat species at 102 sites throughout the Amazon biome. For all possible pairs of sites in both datasets, we estimated the Jaccard pairwise dissimilarity, that is, β‐diversity, considering its three dimensions—taxonomic, phylogenetic and functional—for its two components—turnover (substitution of species) and differences in species richness. The association between dissimilarity measurements and geographical and environment variables was assessed using multiple regressions on distance matrices (MRM).ResultsWe found that turnover and differences in species richness had similar contributions to the taxonomic β‐diversity. However, for phylogenetic and functional β‐diversity, lineages and functions richness differences contribute slightly more than turnover for total β‐diversity. In the lowlands, species, lineages and functions richness differences were slightly higher than turnover for all diversity dimensions. When accounting for all the sites, elevation was the main predictor of phyllostomid bats' taxonomic, phylogenetic and functional turnover. For lowland sites, ecoregions was the main (but relatively weak) predictor associated with all β‐diversity dimensions.Main conclusionsAnalysis of filtering sites according to elevation revealed that species in the Amazonian lowlands are taxonomically and phylogenetically different from species in the Andes, and present taxonomic, phylogenetic and functional redundancy between assemblages. When accounting for the whole range of distribution of bats, results showed the predominant effect of elevation over other geographical and environmental predictors. This indicates that the diversity of good dispersers such as bats is more affected by specialisation along environment and climatic gradients than by geographical barriers throughout the Amazon biome.