Universal Prediction Model Research Articles

Cross-city PM2.5 predictions with recurrent neural network

PM2.5 is inhalable particulate with a diameter less than 2.5 μM that easily enters the lungs and causes diseases and non-accidental death. The generation and dissipation of PM2.5 are strongly affected by a variety of environmental factors, thus the concentration of PM2.5 is presumably predictable with the observations of environmental conditions. This paper used multi-year meteorological and PM2.5 concentration data across multiple megacities in China (Beijing, Chengdu, and Shenyang) and sought for a universal predictive model. Our results showed that data-driven machine-learning model was able to not only capture PM2.5 dynamics at the city where the model was trained but also could be generalized to predict PM2.5 concentrations over other cities. Therefore, the modeling results indicated a universally existing predictive relationship between PM2.5 source-sink dynamics and the environmental drivers.

Bankroto prognozavimo modeliai Europos bankų sektoriui

The global financial crisis has hit the global business and society hard and has caused a mistrust in the monitoring and control of the banking sector a decade ago. The distrust encouraged supervisory authorities to reinforce the bankruptcy preventive measures, that are being developed to prevent or mitigate the negative effects of the new crisis. The development and permanent application of reliable bankruptcy prediction models are the part of such preventive measures. The aim of the study was to create a universal straightforward bankruptcy prediction model for the European banking sector, based on the analysis of financial ratios. The study examines the scholarly literature and deploys the theoretical methods of comparative analysis, critical evaluation, systematisation, generalisation. The empirical research involves expert assessment method, analysis of financial statements, data collection, structuring and modification, modelling: binary logistic regression, correlation, graphical representation of data, ROC curve. Research and results: based on selected 24 variables (ratios), five binary logistic regression models were created. The accuracy was tested on the ROC curve, together with the comparison of II type errors, which were made while predicting bank’s failure 1 year and 2 years before the distress. Main findings were: when models where validated and their II type errors were compared, the conclusion was made that model Z5 is the most suitable for predicting bankruptcy for the European banking sector. The accuracy to identify failing banks 1 year before the distress was 78.57%.
 JEL: F 37, G2, G3

Prediction and Comparison of Shell Condensers With Straight or Helical Channels for Underwater Vehicles

The shell condenser is one of the key components of underwater vehicles. To study its thermal performance and to design a more efficient structure, a computational model is generated to simulate condensation inside straight and helical channels. The model combines empirical correlations and a MATLAB-based iterative algorithm. The vapor quality is used as a sign of the degree of condensation. Three calculation models are compared, and the optimal model is verified by a comparison of simulated results and available experimental data. Several cases are designed to reveal the effects of various inlet conditions and the diameter-over-radius (Dh/R) ratio. The results show that the inlet temperature and mass rate significantly affect the flow and heat transfer in the condensation process, the heat transfer capabilities of the helical channels are much better than that of the straight channel, and both the heat transfer coefficient and total pressure drop increase with the decrease of Dh/R. This study may provide a useful reference for performance prediction and structural design of shell condensers used for underwater vehicles and may provide a relatively universal prediction model for condensation in channels.

Internationalization and failure risk patterns

PurposeThis paper aims to find out what kind of export and failure risk patterns exist among young Estonian manufacturing exporters and explore their interlinkages.Design/methodology/approachThe sample consisted of 208 young Estonian manufacturing exporters. Based on internationalization literature, export patterns were detected with a consecutive three-stage clustering of export sales share from total sales, outside-Europe sales share from export sales and number of target markets, while failure risk patterns were detected by clustering failure probabilities obtained from a universal prediction model. The interconnection of export patterns with financial ratios and failure risk patterns was studied with statistical tests.FindingsSix main internationalization patterns existed. In all, 49 per cent of firms exported to a single European market and their export share was constantly very low, while even most of the firms with high export shares (39 per cent of the sample) were also active on one European market. In terms of failure risk patterns, 49 per cent of firms had constantly very low failure risk, while 51 per cent of firms had medium risk. Higher export engagement did not lead to better financial performance or lower failure risk.Originality/valueThis study is the first to find out if firms following different export patterns are also characterized by specific financial performance and failure risk. In addition, studies encompassing young exporters’ specific target markets and failure risk development are rare. While exporters’ and non-exporters’ financial performance differences have been frequently documented in favor of the former, this study found no such differences for different types of young exporters.

Accurate prediction of surface tension of refrigerants

Surface tension of refrigerants is an important parameter for design problems involving the use of refrigerants. There is no universal and general model for prediction of surface tension of these compounds at different conditions. Hence, developing simple, accurate and general models for prediction of surface tension of these compounds is of great importance. In this work two computer based models named PSO-RBF and Hybrid-ANFIS were developed for prediction of surface tension of refrigerants. A number of 2879 experimental data covering wide ranges of influencing parameters were utilized to develop the models. The accuracy of predictions of the developed models was examined by using different statistical quality measure approaches as well as comparing their results with the predictions of literature models. Results showed that the developed models present accurate predictions and are superior to the literature models. The predictions of PSO-RBF model were also better than Hybrid-ANFIS model.

Can Machine Learning Methods Produce Accurate and Easy-to-use Prediction Models of 30-day Complications and Mortality After Knee or Hip Arthroplasty?

Existing universal and procedure-specific surgical risk prediction models of death and major complications after elective total joint arthroplasty (TJA) have limitations including poor transparency, poor to modest accuracy, and insufficient validation to establish performance across diverse settings. Thus, the need remains for accurate and validated prediction models for use in preoperative management, informed consent, shared decision-making, and risk adjustment for reimbursement. The purpose of this study was to use machine learning methods and large national databases to develop and validate (both internally and externally) parsimonious risk-prediction models for mortality and complications after TJA. Preoperative demographic and clinical variables from all 107,792 nonemergent primary THAs and TKAs in the 2013 to 2014 American College of Surgeons-National Surgical Quality Improvement Program (ACS-NSQIP) were evaluated as predictors of 30-day death and major complications. The NSQIP database was chosen for its high-quality data on important outcomes and rich characterization of preoperative demographic and clinical predictors for demographically and geographically diverse patients. Least absolute shrinkage and selection operator (LASSO) regression, a type of machine learning that optimizes accuracy and parsimony, was used for model development. Tenfold validation was used to produce C-statistics, a measure of how well models discriminate patients who experience an outcome from those who do not. External validation, which evaluates the generalizability of the models to new data sources and patient groups, was accomplished using data from the Veterans Affairs Surgical Quality Improvement Program (VASQIP). Models previously developed from VASQIP data were also externally validated using NSQIP data to examine the generalizability of their performance with a different group of patients outside the VASQIP context. The models, developed using LASSO regression with diverse clinical (for example, American Society of Anesthesiologists classification, comorbidities) and demographic (for example, age, gender) inputs, had good accuracy in terms of discriminating the likelihood a patient would experience, within 30 days of arthroplasty, a renal complication (C-statistic, 0.78; 95% confidence interval [CI], 0.76-0.80), death (0.73; 95% CI, 0.70-0.76), or a cardiac complication (0.73; 95% CI, 0.71-0.75) from one who would not. By contrast, the models demonstrated poor accuracy for venous thromboembolism (C-statistic, 0.61; 95% CI, 0.60-0.62) and any complication (C-statistic, 0.64; 95% CI, 0.63-0.65). External validation of the NSQIP- derived models using VASQIP data found them to be robust in terms of predictions about mortality and cardiac complications, but not for predicting renal complications. Models previously developed with VASQIP data had poor accuracy when externally validated with NSQIP data, suggesting they should not be used outside the context of the Veterans Health Administration. Moderately accurate predictive models of 30-day mortality and cardiac complications after elective primary TJA were developed as well as internally and externally validated. To our knowledge, these are the most accurate and rigorously validated TJA-specific prediction models currently available (http://med.stanford.edu/s-spire/Resources/clinical-tools-.html). Methods to improve these models, including the addition of nonstandard inputs such as natural language processing of preoperative clinical progress notes or radiographs, should be pursued as should the development and validation of models to predict longer term improvements in pain and function. Level III, diagnostic study.

An advanced approach for estimation of PCU values on undivided urban roads under heterogeneous traffic conditions

ABSTRACTMajority of the existing studies on passenger car unit (PCU) estimation are limited to inter-urban and divided urban roads. The characteristics of traffic on undivided urban roads are completely different from those on divided roads. This study aims to estimate the PCU values on undivided urban roads with varying traffic conditions. Estimation of PCU requires speed information for individual vehicle categories. Hence, a universal Kriging-based speed prediction model was developed and utilized for PCU estimation. The model was constructed based on the classified traffic volume and speed data that were collected by videography method on undivided urban road segments in different cities in India. The proposed model was utilized to study the effects of traffic volume, traffic composition, and carriageway width on PCU. An alternative approach of ‘Stream Equivalency Factor’ has also been suggested in this paper. Use of this approach can simplify the process of determining the homogeneous equivalent of a mixed traffic flow.

Gas Flow at the Ultra-nanoscale: Universal Predictive Model and Validation in Nanochannels of Ångstrom-Level Resolution.

Gas transport across nanoscale pores is determinant in molecular exchange in living organisms as well as in a broad spectrum of technologies. Here, we report an unprecedented theoretical and experimental analysis of gas transport in a consistent set of confining nanochannels ranging in size from the ultra-nanoscale to the sub-microscale. A generally applicable theoretical approach quantitatively predicting confined gas flow in the Knudsen and transition regime was developed. Unlike current theories, specifically designed for very simple channel geometries, our approach can be applied to virtually all geometries, for which the probability distribution of path lengths for particle-interface collisions can be computed, either analytically or by numerical simulations. To generate a much needed benchmark experimental model, we manufactured extremely reproducible membranes with two-dimensional nanochannels. Channel sizes ranged from 2.5 to 250 nm, and angstrom level of size control and interface tolerances were achieved using leading-edge nanofabrication techniques. We then measured gas flow in the Knudsen number range from 0.2 to 20. Excellent agreement between theoretical predictions and experimental data was found, demonstrating the validity and potential of our approach.

Surrogate model–based optimal feed-forward control for dimensional-variation reduction in composite parts' assembly processes

ABSTRACTDimension control and variation reduction are vital for composite parts' assembly processes. Due to the nonlinear properties of composites, physics-based models cannot accurately and efficiently approximate the assembly processes. In addition, conventional robust parameter design (RPD) and statistical process control (SPC) cannot actively compensate for dimensional errors or prevent defects. This article proposes a surrogate model–based optimal feed-forward control strategy for dimensional-variation reduction and defect prevention in the assembly of composite parts. The objective is accomplished by (i) developing a grouped Latin hypercube sampling approach tailored to the problem; (ii) adopting a universal Kriging model for dimensional prediction and then embedding the model into an optimal feed-forward control algorithm; and (iii) conducting a multiobjective optimization to determine the control actions. A case study reveals that the developed methodology can effectively reduce the mean and standard deviation of dimensional deviations for the assembly of composite parts.

Development of a universal blast-induced ground vibration prediction model for Jharia coalfields

Development of a universal blast-induced ground vibration prediction model for Jharia coalfields

A Study of GIS-based Rural Land Approval Agglomeration and Economic Growth in China

The No.1 Document of the Party Central Committee put forward to carry out a comprehensive rural land approval registration certification work in 2013. In response to the call, the Ministry of Agriculture accelerate the development of rural land approval registration certification work, to further stabilize the land contract relations, promote land transfer and scale development, add new energy for the agriculture development. By 2016, more than 90% of the country had already carried out land approval work, so this paper employs the rural land approval data to study economy growth on the role of rural land approval agglomeration in most provinces of China in 2016. With applying the geographic information systems in the application, forecasting the model of rural land approval and economic growth is the inevitable trend in development. This article in view of spatial information systems researching in the rural land approval distribution is to make prediction of the toughest task. The results show that rural land approval agglomeration has a significant positive role in the regional economic growth in different models divided by two groups. Then we use universal Kriging prediction model to analysis and validate the correction.

A novel passenger flow prediction model using deep learning methods

Currently, deep learning has been successfully applied in many fields and achieved amazing results. Meanwhile, big data has revolutionized the transportation industry over the past several years. These two hot topics have inspired us to reconsider the traditional issue of passenger flow prediction. Asa special structure of deep neural network (DNN), an autoencoder can deeply and abstractly extract the nonlinear features embedded in the input without any labels. By exploiting its remarkable capabilities, a novel hourly passenger flow prediction model using deep learning methods is proposed in this paper. Temporal features including the day of a week, the hour of a day, and holidays, the scenario features including inbound and outbound, and tickets and cards, and the passenger flow features including the previous average passenger flow and real-time passenger flow, are defined as the input features. These features are combined and trained as different stacked autoencoders (SAE) in the first stage. Then, the pre-trained SAE are further used to initialize the supervised DNN with the real-time passenger flow as the label data in the second stage. The hybrid model (SAE-DNN) is applied and evaluated with a case study of passenger flow prediction for four bus rapid transit (BRT) stations of Xiamen in the third stage. The experimental results show that the proposed method has the capability to provide a more accurate and universal passenger flow prediction model for different BRT stations with different passenger flow profiles.

National-scale exposure prediction for long-term concentrations of particulate matter and nitrogen dioxide in South Korea

The limited spatial coverage of the air pollution data available from regulatory air quality monitoring networks hampers national-scale epidemiological studies of air pollution. The present study aimed to develop a national-scale exposure prediction model for estimating annual average concentrations of PM10 and NO2 at residences in South Korea using regulatory monitoring data for 2010. Using hourly measurements of PM10 and NO2 at 277 regulatory monitoring sites, we calculated the annual average concentrations at each site. We also computed 322 geographic variables in order to represent plausible local and regional pollution sources. Using these data, we developed universal kriging models, including three summary predictors estimated by partial least squares (PLS). The model performance was evaluated with fivefold cross-validation. In sensitivity analyses, we compared our approach with two alternative approaches, which added regional interactions and replaced the PLS predictors with up to ten selected variables. Finally, we predicted the annual average concentrations of PM10 and NO2 at 83,463 centroids of residential census output areas in South Korea to investigate the population exposure to these pollutants and to compare the exposure levels between monitored and unmonitored areas. The means of the annual average concentrations of PM10 and NO2 for 2010, across regulatory monitoring sites in South Korea, were 51.63 μg/m3 (SD = 8.58) and 25.64 ppb (11.05), respectively. The universal kriging exposure prediction models yielded cross-validated R2s of 0.45 and 0.82 for PM10 and NO2, respectively. Compared to our model, the two alternative approaches gave consistent or worse performances. Population exposure levels in unmonitored areas were lower than in monitored areas. This is the first study that focused on developing a national-scale point wise exposure prediction approach in South Korea, which will allow national exposure assessments and epidemiological research to answer policy-related questions and to draw comparisons among different countries.

Determination of the Apparent Carbon Oxidation Reaction Order by a Microfluidized Bed and Its Application to Kinetic Models

During carbon oxidation, the apparent reaction order n ranges from 0 to 1. Considering the importance of the value of n to the accuracy of char combustion rate prediction models and the error caused by subjective selections of its value, this study proposes a fast, simple, and accurate method to determine n, the results of which can be applied to improve the accuracy of kinetic models. In the kinetic control region, the apparent carbon oxidation reaction order n was first determined using a microfluidized bed (MFB). As an example of its application, the value of n thus determined was applied to a universal prediction model of the char combustion rate; the model was further developed using the shrinking-core model to predict the combustion rate of graphite during the entire reaction process. The result shows that the apparent reaction order is close to zero (0.178). The accuracy of predictions using the aforementioned modified model was found to be improved, and the accuracy of the model, as further developed, is nearly unaffected, indicating that this method of determining n has a broader application potential.

A Universal Predictive Model for Dose Fall-off in Multileaf Collimator–based Stereotactic Brain Radiosurgery

A Universal Predictive Model for Dose Fall-off in Multileaf Collimator–based Stereotactic Brain Radiosurgery

PS1 - 189 A Universal Predictive Model for Dose Fall-Off in MLC-Based Stereotactic Brain Radiosurgery

Predictive modeling of dose fall-off in radiosurgery could assist in clinical decision-making when prescribing a treatment plan with minimized toxicity risk. The purpose of this study is to develop a predictive dose fall-off model. Materials/Methods: We retrospectively reviewed treatment plans from 257 patients (365 lesions) with total doses ranging from 20 to 35Gy in 5 fractions. For each plan, we measured both total volume of the external contour (EXT) and BrainMinusPTV (BMP) receiving P=20% to P=80% of the prescription dose. The model has form y=Fa(PTV)b+/-delta. y=volume of EXT or BMP (cc’s); a and b are curve-fitting coefficients; PTV=total planning target volume (cc’s); F is an adjustment factor (>1) to account for number of targets; delta is the 95% prediction band. F, a, b, and delta were modeled such that dose-fall can be forecast for any PTV and dose level. Results: The model coefficients were as follows: Coefficient EXT BMP a 19927(100×P)exp(-2) 17122(100×P)exp(-2) b 0.42(100×P)exp(0.17) 0.63 F -0.0156×(100×P)+2.5517 delta 384467×(100×P)exp(-2.3159) The table can be used to determine the model for any P from 20% to 80%. Example: the EXT receiving 50%, P=0.5, a=8.0, b=0.82, F=1.8, delta=45. Thus, EXT-50=8(PTV0.82) or 1.8×8(PTV0.82) for 1-3 or >3 targets, respectively,+/-45cc’s. The model was verified against published values of dose fall-off from linacs. Conclusion: A predictive dose fall-off model was generated for linac-based radiosurgery. The model can be used for quality assurance or for inter-institutional comparisons. Ongoing work is being conducted to extend the model to a SRS cones system.

Towards building a universal defect prediction model with rank transformed predictors

Software defects can lead to undesired results. Correcting defects costs 50 % to 75 % of the total software development budgets. To predict defective files, a prediction model must be built with pr...

Diagnosis-specific readmission risk prediction using electronic health data: a retrospective cohort study.

BackgroundReadmissions after hospital discharge are a common occurrence and are costly for both hospitals and patients. Previous attempts to create universal risk prediction models for readmission have not met with success. In this study we leveraged a comprehensive electronic health record to create readmission-risk models that were institution- and patient- specific in an attempt to improve our ability to predict readmission.MethodsThis is a retrospective cohort study performed at a large midwestern tertiary care medical center. All patients with a primary discharge diagnosis of congestive heart failure, acute myocardial infarction or pneumonia over a two-year time period were included in the analysis.The main outcome was 30-day readmission. Demographic, comorbidity, laboratory, and medication data were collected on all patients from a comprehensive information warehouse. Using multivariable analysis with stepwise removal we created three risk disease-specific risk prediction models and a combined model. These models were then validated on separate cohorts.Results3572 patients were included in the derivation cohort. Overall there was a 16.2% readmission rate. The acute myocardial infarction and pneumonia readmission-risk models performed well on a random sample validation cohort (AUC range 0.73 to 0.76) but less well on a historical validation cohort (AUC 0.66 for both). The congestive heart failure model performed poorly on both validation cohorts (AUC 0.63 and 0.64).ConclusionsThe readmission-risk models for acute myocardial infarction and pneumonia validated well on a contemporary cohort, but not as well on a historical cohort, suggesting that models such as these need to be continuously trained and adjusted to respond to local trends. The poor performance of the congestive heart failure model may suggest that for chronic disease conditions social and behavioral variables are of greater importance and improved documentation of these variables within the electronic health record should be encouraged.

OC-0507: A universal predictive model for diode correction factors in small radiation fields

OC-0507: A universal predictive model for diode correction factors in small radiation fields

A Universal Model to Predict Roadheaders’ Cutting Performance / Uniwersalny Model Do Prognozowania Postępu Prac Kombajnów Do Drążenia Tuneli

Abstract The paper intends to generate a universal model to predict the performance of roadheaders for all kinds of rock formations. In this regard, we first take into account the outcomes of previous attempts to explore the performance of roadheaders in Tabas Coal Mine project (the largest and fully mechanized coal mine in Iran). During those investigations, rock mass brittleness index (RMBI) was defined in order to relate the intact and rock mass characteristics to machine performance. The statistical analysis of data acquired from Tabas field demonstrated that RMBI was highly correlated to instantaneous cutting rate (ICR) of roadheaders (R² = 0.92). With the aim to construct a universal model for predicting the roadheader performance, we have now tried to establish a database consisting measured cutting rate of roadheaders as well as the data gathered from field studies of Tabas Coal Mine project and Besiktas, Kurucesme, Baltalimani, Eyup and Halic tunnels in Turkey. A broad modeling and analysis found a fair relationship, resulting in a new universal predictive model to predict the cutting rate of roadheaders with correlation of 0.73 (R² = 0.73). The application of local and universal models at Tabas Coal Mine showed a remarkable difference between measured and predicted ICR. The mean relative error of 0.359% was found with universal model but it represented lower value (mean relative error of 0.100%) while using local model. It can thus be concluded that instead of generating a universal model, separate localized models for different ground and machine conditions should be developed to improve the accuracy and reliability of the performance prediction models.