Set Of Input Variables Research Articles

Performance Comparison Between SURPAS and ACS NSQIP Surgical Risk Calculator in Pulmonary Resection

Accurate preoperative risk assessment is critical for informed decision making. The Surgical Risk Preoperative Assessment System (SURPAS) and the National Surgical Quality Improvement Program (NSQIP) Surgical Risk Calculator (SRC) predict risks of common postoperative complications. This study compares observed and predicted outcomes after pulmonary resection between SURPAS and NSQIP SRC. Between January 2016 and December 2018, 2514 patients underwent pulmonary resection and were included. We entered the requisite patient demographics, preoperative risk factors, and procedural details into the online NSQIP SRC and SURPAS formulas. Performance of the prediction models was assessed by discrimination and calibration. No statistically significant differences were found between the 2 models in discrimination performance for 30-day mortality, urinary tract infection, readmission, and discharge to a nursing or rehabilitation facility. The ability to discriminate between a patient who will develop a complication and a patient who will not was statistically indistinguishable between NSQIP and SURPAS, except for renal failure. With a C index closer to 1.0, the NSQIP performed significantly better than the SURPAS SRC in discriminating risk of renal failure (C index, 0.798 vs 0.694; P= .003). The calibration curves of predicted and observed risk for each model demonstrate similar performance with a tendency toward overestimation of risk, apart from renal failure. Overall, SURPAS and NSQIP SRC performed similarly in predicting outcomes for pulmonary resections in this large, single-center validation study with moderate to good discrimination of outcomes. Notably, SURPAS uses a smaller set of input variables to generate the preoperative risk assessment. The addition of thoracic-specific input variables may improve performance.

Predicting the dynamic modulus of hot mix asphalt mixtures using bagged trees ensemble

Bagged trees ensemble, a machine-learning algorithm, was used for the first time to predict the dynamic modulus |E*| of dense-graded HMA mixtures with success. This algorithm contained a similar set of input variables as the NCHRP 1-40D and Hirsch models of |E*| and was implemented in the statistics toolbox of Matlab R2018b. It was trained and tested on 1656 |E*| values from 93 typical South Carolina HMA mixtures. Of the tested mixtures, 61% contained RAP. For comparison purposes, the performance of bagged trees was compared with two already published ANNs in the literature. The results showed that: 1) Coupling VMA and Gb* input variables had the most significant influence on |E*|. 2) Bagged trees exhibited a significantly better prediction accuracy than the Bari-Witczak and Hirsch regression models with an R2 greater than 0.94 on testing. 3) Bagged trees performed better than two already published ANN models of |E*|. In conclusion, bagged trees is a useful tool for estimating |E*| of HMA mixtures.

COMBINING ENVIRONMENTAL AND LANDSAT ANALYSIS READY DATA FOR VEGETATION MAPPING: A CASE STUDY IN THE BRAZILIAN SAVANNA BIOME

Abstract. The Cerrado biome in Brazil covers approximately 24% of the country. It is one of the richest and most diverse savannas in the world, with 23 vegetation types (physiognomies) consisting mostly of tropical savannas, grasslands, forests and dry forests. It is considered as one of the global hotspots of biodiversity because of the high level of endemism and rapid loss of its original habitat. This work aims to analyze the potential of Landsat Analysis Ready Data (ARD) in combination with different environmental data to classify the vegetation in the Cerrado in two different hierarchical levels. Here we present results of a pixel-based modelling exercise, in which field data were combined with a set of input variables using a Random Forest classification approach. On the first hierarchical level, with the three classes savanna, grasslands and forest, our model results reached f1-scores of 0.86, 0.87 and 0.85 leading to an overall accuracy of 0.86. In the second hierarchical level we differentiated a total of 12 vegetation physiognomies with an overall accuracy of 0.77.

A Primer on Machine Learning.

A Primer on Machine Learning.

Classifying dual role variables in DEA: The case of VRS

Data envelopment analysis is a mathematical programming methodology commonly used to evaluate the relative efficiencies of each of a set of decision-making units in terms of a selected set of input and output variables. In many situations, however, it can happen that certain variables can play roles of either inputs or outputs. This dual-role or classification concept was examined in earlier research for settings where the production possibility set was defined in terms of a constant returns to scale (CRS) technology. The current article extends the classification models to those cases where a variable returns to scale (VRS) technology is present. We show that if one simply adds a free-in-sign variable to the CRS model, the resulting VRS model can be unbounded. To alleviate this circumstance, we develop an altered version of the conventional VRS model which is shown to be feasible and bounded. We provide numerical examples to demonstrate the applicability and strength of the proposed methodology.

A comparative analysis of statistical and machine learning techniques for mapping the spatial distribution of groundwater salinity in a coastal aquifer

Groundwater salinity in an aquifer system is typically measured through field studies (e.g., groundwater sampling, and direct current resistivity method). The field-based measurements are costly and time-consuming when they are applied over a large domain. In this study, a methodology was developed and evaluated based on available hydrogeology and hydrometeorology data and statistical and machine learning techniques to map the groundwater salinity in the southern coastal aquifer of the Caspian Sea. First, variables affecting groundwater salinity (aquifer transmissivity, distance from the sea, the mean annual precipitation, the mean annual evaporation, elevation, and the depth to the water table) were determined, and the dataset was randomly divided into three subsets of training, testing, and verification. Next, the relationship between groundwater salinity and its controlling factors was established using three methods namely extreme gradient boosting (EGB), deep neural network (DNN), and multiple linear regression (MLR), and the models were evaluated by comparing the measured values and the predicted values using the statistical criteria (R-squared, Nash–Sutcliffe efficiency (NSE), and normalized root-mean-square deviation (NRMSD)). Finally, the optimum model was applied to the set of known input variables to map the spatial variation of the groundwater salinity across the entire southern coastal plain of the Caspian Sea, and the final map was verified using the verification subset. Results showed that consideration should be given to the EGB method, considering its higher performance on the testing subset (R-squared 0.89, NSE 0.87, NRMSD 0.45). In-depth analysis of the variables showed that the aquifer transmissivity is the most crucial parameter affecting groundwater salinity in the region. The adopted approach could potentially be used for groundwater management purposes in the study area and similar settings elsewhere.

A Priori Sub-grid Modelling Using Artificial Neural Networks

This paper presents results of Artificial Neural Networks (ANN) applications to sub-grid Large Eddy Simulation (LES) model. The training data for the ANN is provided by simulation of Homogeneous Isotropic Turbulence at different Reynolds numbers. The results show that the correlation coefficients are superior to other sub-grid models, using a similar set of input variables. As the ANN model extrapolates to larger Reynolds, the correlation coefficient decreases. However, it remains higher than other sub-grid approaches, and suggest that the combined LES-ANN methodology can potentially be used as a sub-grid model at realistic Reynolds numbers. Models derived from Homogeneous Isotropic Turbulence can also be used in different simple flows and provide relatively good agreement.

Experimental investigation of the behaviour of a DI diesel engine fuelled with biodiesel/diesel blends having effect of raw biogas at different operating responses

The ever-growing consumption and rapid exhaustion of fossil fuels along with environmental concern continue to be an important issue for researchers around the world. The quest for improving direct injection (DI) diesel engine performance and reducing emissions has led researchers to explore various aspects of alternate sources of fuels and also have a possibility of dual fuel mode operation. These DI diesel engines use conventional liquid diesel fuel supplemented with gaseous fuels. The present contribution involves, experimental investigation as well as the implementation of response surface methodology (RSM) technique to evaluate the performance of DI diesel engine using biodiesel blends of polanga biodiesel and alcohol along with dual fueling of the engine when gaseous fuel i.e. biogas partially replaces the liquid diesel in the engine. The input parameters considered are types of fuel, speed of the engine, air flow rate and engine load whereas brake thermal efficiency (BTE), volumetric efficiency (VE), and exhaust gas temperature (EGT) along with emission parameters carbon monoxide (CO), (nitrogen oxides) NOx and hydrocarbon (HC) were considered to be the resulting output parameters. The overall good engine performance has been observed at an optimum setting of engine input variables through RSM which is 1495 RPM, 5.8 kg load, 48 mm air flow rate and polanga biodiesel with 10% biogas. The corresponding optimum engine responses of VE, BTE, EGT, NOx, HC, and CO has been observed as 68.92%, 23.62%, 172 oK, 966 ppm volume, 34.65 ppm volume and 0.0876 ppm volume respectively.

Deep reinforcement learning to optimise indoor temperature control and heating energy consumption in buildings

In this work, Deep Reinforcement Learning (DRL) is implemented to control the supply water temperature setpoint to terminal units of a heating system. The experiment was carried out for an office building in an integrated simulation environment. A sensitivity analysis is carried out on relevant hyperparameters to identify their optimal configuration. Moreover, two sets of input variables were considered for assessing their impact on the adaptability capabilities of the DRL controller. In this context a static and dynamic deployment of the DRL controller is performed. The trained control agent is tested for four different scenarios to determine its adaptability to the variation of forcing variables such as weather conditions, occupant presence patterns and different indoor temperature setpoint requirements. The performance of the agent is evaluated against a reference controller that implements a combination of rule-based and climatic-based logics. As a result, when the set of variables are adequately selected a heating energy saving ranging between 5 and 12% is obtained with an enhanced indoor temperature control with both static and dynamic deployment. Eventually the study proves that if the set of input variables are not carefully selected a dynamic deployment is strictly required for obtaining good performance.

Multiresponse optimization of a multistage manufacturing process using a patient rule induction method

AbstractMost manufacturing industries produce products through a series of sequential stages, known as a multistage process. In a multistage process, each stage affects the stage that follows, and the process often has multiple response variables. In this paper, we suggest a new procedure for optimizing a multistage process with multiple response variables. Our method searches for an optimal setting of input variables directly from operational data according to a patient rule induction method (PRIM) to maximize a desirability function, to which multiple response variables are converted. The proposed method is explained by a step‐by‐step procedure using a steel manufacturing process as an example. The results of the steel manufacturing process optimization show that the proposed method finds the optimal settings of input variables and outperforms the other PRIM‐based methods.

Artificial intelligence and organ transplantation: challenges and expectations.

Classifiers based on artificial intelligence have emerged in all areas of medicine. Although very subtle, many decisions in organ transplantation can now be addressed in a more concisely manner with the support of these classifiers. Any aspect of organ transplantation (image processing, prediction of results, diagnostic proposals, therapeutic algorithms or precision treatments) consists of a set of input variables and a set of output variables. Artificial intelligence classifiers differ in the way they establish relationships between the input variables, how they select the data groups to train patterns and how they are able to predict the possible options of the output variables. There are hundreds of classifiers to achieve this goal. The most appropriate classifiers to address the different aspects of organ transplantation are Artificial Neural Networks, Decision Tree classifiers, Random Forest, and Naïve Bayes classification models. There are hundreds of examples of the usefulness of artificial intelligence in organ transplantation, especially in image processing, organ allocation, D-R matching, precision pathology, real-time immunosuppression, transplant oncology, and predictive analysis. In the coming years, clinical transplant experts will increasingly use Deep Learning-based models to support their decisions, specially in those cases where subjectivity is common.

Automated detector simulation and reconstruction parametrization using machine learning

Rapidly applying the effects of detector response to physics objects (e.g. electrons, muons, showers of particles) is essential in high energy physics. Currently available tools for the transformation from truth-level physics objects to reconstructed detector-level physics objects involve manually defining resolution functions. These resolution functions are typically derived in bins of variables that are correlated with the resolution (e.g. pseudorapidity and transverse momentum). This process is time consuming, requires manual updates when detector conditions change, and can miss important correlations. Machine learning offers a way to automate the process of building these truth-to-reconstructed object transformations and can capture complex correlation for any given set of input variables. Such machine learning algorithms, with sufficient optimization, could have a wide range of applications: improving phenomenological studies by using a better detector representation, allowing for more efficient production of Geant4 simulation by only simulating events within an interesting part of phase space, and studies on future experimental sensitivity to new physics.

Classifier Neural Network Models Predict Relativistic Electron Events at Geosynchronous Orbit Better than Multiple Regression or ARMAX Models

AbstractTo find the best method of predicting when daily relativistic electron flux (>2 MeV) will rise at geosynchronous orbit, we compare model predictive success rates (true positive rate or TPR) for multiple regression, ARMAX, logistic regression, a feed‐forward multilayer perceptron (MLP), and a recurrent neural network (RNN) model. We use only those days on which flux could rise, removing days when flux is already high from the data set. We explore three input variable sets: (1) ground‐based data (Kp, Dst, and sunspot number), (2) a full set of easily available solar wind and interplanetary magnetic field parameters (|B|, Bz, V, N, P, Ey, Kp, Dst, and sunspot number, and (3) this full set with the addition of previous day's flux. Despite high validation correlations in the multiple regression and ARMAX predictions, these regression models had low predictive ability (TPR < 45%) and are not recommended for use. The three classifier model types (logistic regression, MLP, and RNN) performed better (TPR: 50.8–74.6%). These rates were increased further if the cost of missing an event was set at 4 times that of predicting an event that did not happen (TPR: 73.1–89.6%). The area under the receiver operating characteristic curves did not, for the most part, differ between the classifier models (logistic, MLP, and RNN), indicating that any of the three could be used to discriminate between events and nonevents, but validation suggests a full RNN model performs best. The addition of previous day's flux as a predictor provided only a slight advantage.

Performance comparison of nonlinear and linear regression algorithms coupled with different attribute selection methods for quantitative structure - retention relationships modelling in micellar liquid chromatography

In micellar liquid chromatography (MLC), the addition of a surfactant to the mobile phase in excess is accompanied by an alteration of its solubilising capacity and a change in the stationary phase's properties. As an implication, the prediction of the analytes’ retention in MLC mode becomes a challenging task. Mixed Quantitative Structure – Retention Relationships (QSRR) modelling represents a powerful tool for estimating the analytes’ retention.This study compares 48 successfully developed mixed QSRR models with respect to their ability to predict retention of aripiprazole and its five impurities from molecular structures and factors that describe the Brij - acetonitrile system. The development of the models was based on an automatic combining of six attribute (feature) selection methods with eight predictive algorithms and the optimization of hyper-parameters. The feature selection methods included Principal Component Analysis (PCA), Non-negative Matrix Factorization (NMF), ReliefF, Multiple Linear Regression (MLR), Mutual Info and F-Regression. The series of investigated predictive algorithms comprised Linear Regressions (LR), Ridge Regression, Lasso Regression, Artificial Neural Networks (ANN), Support Vector Regression (SVR), Random Forest (RF), Gradient Boosted Trees (GBT) and K-Nearest neighbourhood (k-NN).A sufficient amount of data for building the model (78 cases in total) was provided by conducting 13 experiments for each of the 6 analytes and collecting the target responses afterwards. Different experimental settings were established by varying the values of the concentration of Brij L23, pH of the aqueous phase and acetonitrile content in the mobile phase according to the Box-Behnken design. In addition to the chromatographic parameters, the pool of independent variables was expanded by 27 molecular descriptors from all major groups (physicochemical, quantum chemical, topological and spatial structural descriptors). The best model was chosen by taking into consideration the Root Mean Square Error (RMSE) and cross-validation (CV) correlation coefficient (Q2) values.Interestingly, the comparative analysis indicated that a change in the set of input variables had a minor impact on the performance of the final models. On the other hand, different regression algorithms showed great diversity in the ability to learn patterns conserved in the data. In this regard, testing many regression algorithms is necessary in order to find the most suitable technique for model building. In the specific case, GBT-based models have demonstrated the best ability to predict the retention factor in the MLC mode. Steric factors and dipole-dipole interactions have proven to be relevant to the observed retention behaviour. This study, although being of a smaller scale, is a most promising starting point for comprehensive MLC retention prediction.

FUZZY INFERENCE SYSTEMS BASE ON POLYNOMIALCONSEQUENTS OF FUZZY RULES

Various fuzzy inference systems that operate on the basis of polynomial consequents of fuzzy rules. As well as inference methods for such systems, in particular, Takagi-Sugeno fuzzy inference systems, their differences from other popular fuzzy systems, such as Mamdani systems, etc., are considered. The attention is focused on the features of the functioning of such systems both in the construction of elementary fuzzy systems. The Systems for which the calculation of the general logical conclusion involves intermediate levels of logical inference with many hierarchically interconnected blocks of fuzzy rules. Fuzzy sets of type 2 are considered, the membership index of which is a fuzzy term of the first type. This allows you to take into account the secondary fuzziness of linguistic concepts in the design of intelligent systems based on fuzzy inference. Fuzzy systems of the second type based on Takagi-Sugeno systems and the iterative Karnik-Mendel algorithm are considered to obtain a logical conclusion for fuzzy systems with the interval membership functions of the second type in the antecedents of fuzzy rules. The proposed procedure for lowering the order of fuzzy rules for higher-order Takagi-Sugeno fuzzy systems is described and justified. A fuzzy inference method for higher-order fuzzy systems based on the partition of a set of input variables is proposed. It is proposed to build a separate block of fuzzy rules for each of the input subspaces in the presence of a common polynomial. Which is a higher-order consequent, that reduces the total number of fuzzy rules in blocks.

Data driven models for compressive strength prediction of concrete at high temperatures

The use of data driven models has been shown to be useful for simulating complex engineering processes, when the only information available consists of the data of the process. In this study, four data-driven models, namely multiple linear regression, artificial neural network, adaptive neural fuzzy inference system, and K nearest neighbor models based on collection of 207 laboratory tests, are investigated for compressive strength prediction of concrete at high temperature. In addition for each model, two different sets of input variables are examined: a complete set and a parsimonious set of involved variables. The results obtained are compared with each other and also to the equations of NIST Technical Note standard and demonstrate the suitability of using the data driven models to predict the compressive strength at high temperature. In addition, the results show employing the parsimonious set of input variables is sufficient for the data driven models to make satisfactory results.

Machine learning algorithms applied to a prediction of personal overall thermal comfort using skin temperatures and occupants' heating behavior

Thermal comfort modeling has been of interest in built environment research for decades. Mostly the modeling approaches focused on an average response of a large group of building occupants. Recently, the focus has been shifted towards personal comfort models that predict individuals' thermal comfort responses. Currently, thermal comfort responses are collected from the occupants via survey. This study explored if the thermal comfort of individuals could be predicted using machine learning algorithms while relaying on the set of collected inputs from an experiment. The model was developed using experimental data including collected from a previously performed experiment in the climate chamber. Two different approaches based on the output data (thermal sensation and thermal comfort votes) and five different sets of input variables were explored. The algorithms tested were Support Vector Machine with four different Kernel functions (Linear, Quadratic, Cubic and Gaussian) and Ensemble Algorithms (Boosted trees, Bagged trees and RUSBoosted trees). The combination of occupants' heating behavior with a personal comfort system (PCS), skin temperatures, time and environmental data were used for the development of personal comfort models to predict individuals' thermal preference. The study investigated the novel combination of inputs such as the use of skin temperature and settings of the personalized heating system as parameters in predicting personal thermal comfort. The results showed that personal comfort models among all tested approaches and subjects showed the best median accuracy of 0.84 using RUSBoosted trees. Individually looking, the approach using thermal sensation output produced better prediction accuracy. On the other hand, the models based on inputs that consisted of PCS control behavior and mean and hand skin temperatures produced the best prediction accuracy when assessing all tested algorithms. The main limitation of the study is the number of test subjects, and further recommendation is to perform more experiments.

OPTIMIZATION AND DECISION MAKING STAGES FOR ANALYSIS OF DUAL RESPONSE PROBLEM: MEDIAN-MAD, NSGA-II, TOPSIS

The purpose of this study is to obtain a proper set of input variables for replicated response measured data set by applying dual response strategy in robust framework with multi-objective perspective. The replicated response measures were transformed to dual response by using robust statistics, median (MED) and median absolute deviation (MAD), instead of mean-standard deviation statistics which were very commonly used in many existing studies. A compromise solution of proposed robust dual response model was obtained via a multi-criteria decision making approach since the optimization was achieved in multi-objective point of view. In this study, well-known two methods, NSGA-II and TOPSIS, were preferred for optimization and decision making stages, respectively. Quality of printing ink data set was used an application from the literature. It is seen from the analysis results that the performance of the proposed robust dual response model was encouraging with the most satisfactory input settings.

Multivariate Adaptive Generalized Poisson Regression Spline (MAGPRS) on the number of acute respiratory infection infants

Acute Respiratory Infection (ARI) is an infectious disease of the respiratory tract that affects the structure of the respiratory tract. The ARI is a health problem that should not be ignored because it causes high infant mortality. Therefore, it is important to know the factors that influence the ARI. Generalized Poisson Regression (GPR) is one of method that can handle cases of overdispersion (variance is greater than the mean) or underdispersion (variance is less than the mean). Multivariate Adaptive Regression Spline as a statistical method for fitting the relationship between a set of input variables and dependent variables. This research is the development of the MARS method and GPR namely MAGPRS. The application of the MAGPRS model was carried out in the case of the number of ARI in infants from surabaya health departement 2017. The results showed that the importance of predictor variables in MAGPRS, the variables affecting the number of ARI patients in infants are the percentage of low birth weight (X2), the percentage of unhealthy houses (X5), and the percentage given non-exclusive breastfeeding to infants (X1).

New statistical method identifies cytokines that distinguish stool microbiomes

Regressing an outcome or dependent variable onto a set of input or independent variables allows the analyst to measure associations between the two so that changes in the outcome can be described by and predicted by changes in the inputs. While there are many ways of doing this in classical statistics, where the dependent variable has certain properties (e.g., a scalar, survival time, count), little progress on regression where the dependent variable are microbiome taxa counts has been made that do not impose extremely strict conditions on the data. In this paper, we propose and apply a new regression model combining the Dirichlet-multinomial distribution with recursive partitioning providing a fully non-parametric regression model. This model, called DM-RPart, is applied to cytokine data and microbiome taxa count data and is applicable to any microbiome taxa count/metadata, is automatically fit, and intuitively interpretable. This is a model which can be applied to any microbiome or other compositional data and software (R package HMP) available through the R CRAN website.