Random Neural Network Research Articles

Testing machine learning algorithms for the prediction of depositional fluxes of the radionuclides 7Be, 210Pb and 40K

The monthly depositional fluxes of 7Be, 210Pb and 40K were measured at Malaga, (Southern Spain) from 2005 to 2018. In this work, the depositional fluxes of these radionuclides are investigated and their relations with several atmospheric variables have been studied by applying two popular machine learning methods: Random Forest and Neural Network algorithms. We extensively test different configurations of these algorithms and demonstrate their predictive ability for reproducing depositional fluxes. The models derived with Neural Networks achieve slightly better results, in average, although similar, having into account the uncertainties. The mean Pearson-R coefficients, evaluated with a k-fold cross-validation method, are around 0.85 for the three radionuclides using Neural Network models, while they go down to 0.83, 0.79 and 0.8 for 7Be, 210Pb and 40K, respectively, for the Random Forest models. Additionally, applying the Recursive Feature Elimination technique we determine the variables more correlated with the depositional fluxes of these radionuclides, which elucidates the main dependences of their temporal variability.

Estimating Path Travel Costs in Large-Scale Networks Using Machine-Learning Techniques

The integration of activity-based models (ABMs) with simulation-based dynamic traffic assignment (DTA) is proposed as a prominent alternative to the traditional four-step approach to modeling transportation networks. In DTA, the least-generalized cost path-finding algorithm is executed to route travelers to their optimal paths, while ABM requires travel-cost skims of the activities, regardless of actual paths, to assess various options available to users such as destination and mode choices. These skims can be calculated either by rerunning the shortest-path-finding algorithm or storing the dynamic travel-cost skims generated at the DTA level. Having a large-scale network with multiple user classes, these methods require high memory storage and run time. Therefore, this study presents a heuristic algorithm utilizing machine-learning methods to predict origin–destination (OD) travel distances, times, and costs using historical vehicle trajectory data generated at the DTA level. This study uses an ellipse boundary around the ODs to truncate the network and extract the relevant vehicle sub-trajectories. Extracted sub-trajectories are clustered into with-toll and non-toll groups based on trip costs to estimate travel distance, time, and monetary cost accordingly. This approach can also be used in navigation systems to have a better estimate of arrival times. Numerical results for the greater Chicago network show promising performance in travel-cost measures estimation with improved solution times. To test the scalability of the proposed method for different OD pairs, the prediction of the best ellipse size is carried out using random forest and neural network methods.

Study on Traffic Accident Forecast of Urban Excess Tunnel Considering Missing Data Filling

Urban highway tunnels are frequent accident locations, and predicting and analyzing road conditions after accidents to avoid traffic congestion is a key measure for tunnel traffic operation management. In this paper, 200 traffic accident data from the YingTian Street Tunnel in Nanjing city were analyzed and encoded to extract the main factors affecting tunnel traffic conditions from three aspects: time, traffic flow, and tunnel environment. Next, graph convolution long short-term memory networks were used to predict and fill in missing traffic flow data. Finally, seven independent variables selected by Pearson correlation analysis were input into the constructed BP neural network and random forest model to predict tunnel traffic conditions during accidents and accident duration. Experimental results show that the accuracy of random forest and BP neural networks in predicting traffic flow is 83.39% and 82.94%, respectively, and that the absolute error of the two models in predicting accident duration is 75% and 60% within 25 min, respectively. Both models perform well in predicting traffic conditions, and the random forest models perform better in terms of robustness and generalization in predicting crash duration. The experimental results have important implications for tunnel operation management during accidents.

The Deep Learning Generative Adversarial Random Neural Network in data marketplaces: The digital creative

Generative Adversarial Networks (GANs) have been proposed as a method to generate multiple replicas from an original version combining a Discriminator and a Generator. The main applications of GANs have been the casual generation of audio and video content. GANs, as a neural method that generates populations of individuals, have emulated genetic algorithms based on biologically inspired operators such as mutation, crossover and selection. This article presents the Deep Learning Generative Adversarial Random Neural Network (RNN) with the same features and functionality as a GAN. Furthermore, the presented algorithm is proposed for an application, the Digital Creative, that generates tradeable replicas in a Data Marketplace, such as 1D functions or audio, 2D and 3D images and video content. The RNN Generator creates individuals mapped from a latent space while the GAN Discriminator evaluates them based on the true data distribution. The performance of the Deep Learning Generative Adversarial RNN has been assessed against several input vectors with different dimensions, in addition to 1D functions and 2D images. The presented results are successful: the learning objective of the RNN Generator creates tradeable replicas at low error, whereas the RNN Discriminator learning target identifies unfit individuals.

Depression prediction using machine learning algorithms

Nowadays, depression is one of the most common negative emotions between people. The goal of our paper is to predict whether the individuals has suffered depression according to the provided 23 features. This study classifies the characteristics and studies their impact on depression by random forest, naive bayes, Support Vector Machine (SVM) and neural network. The authors test all those methods on the public dataset, which conclude that random forest and neural network achieve the same top-1 accuracy (which is 86.8%). Furthermore, the authors also conduct the feature correlation to exploit the most important factors. Results show that the most two influential features leading to depression is age and education level, which means the group of teenagers is the easiest members to have depression emotion. The study also analyses how can people sort the situation and decrease the possibility of suffering from depression.

Machine learning and physical based modeling for cardiac hypertrophy

Background and objectivePredicting the long-term expansion and remodeling of the left ventricle in patients is challenging task but it has the potential to be clinically very useful. MethodsIn our study, we present machine learning models based on random forests, gradient boosting, and neural networks, used to track cardiac hypertrophy. We collected data from multiple patients, and then the model was trained using the patient's medical history and present level of cardiac health. We also demonstrate a physical-based model, using the finite element procedure to simulate the development of cardiac hypertrophy. ResultsOur models were used to forecast the evolution of hypertrophy over six years. The machine learning model and finite element model provided similar results. ConclusionsThe finite element model is much slower, but it's more accurate compared to the machine learning model since it's based on physical laws guiding the hypertrophy process. On the other hand, the machine learning model is fast but the results can be less trustworthy in some cases. Both of our models, enable us to monitor the development of the disease. Because of its speed machine learning model is more likely to be used in clinical practice. Further improvements to our machine learning model could be achieved by collecting data from finite element simulations, adding them to the dataset, and retraining the model. This can result in a fast and more accurate model combining the advantages of physical-based and machine learning modeling.

Interpretable machine learning for psychological research: Opportunities and pitfalls.

In recent years, machine learning methods have become increasingly popular prediction methods in psychology. At the same time, psychological researchers are typically not only interested in making predictions about the dependent variable, but also in learning which predictor variables are relevant, how they influence the dependent variable, and which predictors interact with each other. However, most machine learning methods are not directly interpretable. Interpretation techniques that support researchers in describing how the machine learning technique came to its prediction may be a means to this end. We present a variety of interpretation techniques and illustrate the opportunities they provide for interpreting the results of two widely used black box machine learning methods that serve as our examples: random forests and neural networks. At the same time, we illustrate potential pitfalls and risks of misinterpretation that may occur in certain data settings. We show in which way correlated predictors impact interpretations with regard to the relevance or shape of predictor effects and in which situations interaction effects may or may not be detected. We use simulated didactic examples throughout the article, as well as an empirical data set for illustrating an approach to objectify the interpretation of visualizations. We conclude that, when critically reflected, interpretable machine learning techniques may provide useful tools when describing complex psychological relationships. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Prediction of Ureteral Injury During Colorectal Surgery Using Machine Learning.

Ureteral injury (UI) is a rare but devastating complication during colorectal surgery. Ureteral stents may reduce UI but carry risks themselves. Risk predictors for UI could help target the use of stents, but previous efforts have relied on logistic regression (LR), shown moderate accuracy, and used intraoperative variables. We sought to use an emerging approach in predictive analytics, machine learning, to create a model for UI. Patients who underwent colorectal surgery were identified in the National Surgical Quality Improvement Program (NSQIP) database. Patients were split into training, validation, and test sets. The primary outcome was UI. Three machine learning approaches were tested including random forest (RF), gradient boosting (XGB), and neural networks (NN), and compared with traditional LR. Model performance was assessed using area under the curve (AUROC). The data set included 262,923 patients, of whom 1519 (.578%) experienced UI. Of the modeling techniques, XGB performed the best, with an AUROC score of .774 (95% CI .742-.807) compared with .698 (95% CI .664-.733) for LR. Random forest and NN performed similarly with scores of .738 and .763, respectively. Type of procedure, work RVUs, indication for surgery, and mechanical bowel prep showed the strongest influence on model predictions. Machine learning-based models significantly outperformed LR and previous models and showed high accuracy in predicting UI during colorectal surgery. With proper validation, they could be used to support decision making regarding the placement of ureteral stents preoperatively.

Evaluation of the dataset quality in gamma passing rate predictions using machine learning methods.

Gamma passing rate (GPR) predictions using machine learning methods have been explored for treatment verification of radiotherapy plans. However, these methods presented datasets with unbalanced number of plans having different treatment conditions (heterogeneous datasets), such as anatomical sites or dose per fractions, leading to lower model interpretability and prediction performance. We investigated the impact of the dataset composition on GPR binary classification (pass/fail) using random forest (RF), XG-boost, and neural network (NN) models. 945 plans were used to create one reference dataset (randomly assembled) and 24 customized datasets that considered four heterogeneity factors independently (anatomical region, number of arcs, dose per fraction, and treatment unit). 309 predictor features were extracted and calculated from plan parameters, modulation complexity metrics, and radiomic analysis (leave-trajectory maps, 3D dose distributions, and portal dosimetry images). The models' performances were measured using the area under the curve from the receiver operating characteristic (ROC-AUC). Radiomics features for reference models increased ROC-AUC values up to 13%, 15%, and 5% for RF, XG-Boost, and NN, respectively. The datasets with higher heterogeneous conditions presented the lower ROC-AUC values (RF: 0.72 ± 0.11, XG-Boost: 0.67 ± 0.1, NN: 0.89 ± 0.05) compared to models with less heterogeneous treatment conditions (RF: 0.88 ± 0.06, XG-Boost: 0.89 ± 0.07, NN: 0.98 ± 0.01). The ten most important features for each heterogeneity dataset group demonstrated their correlation with the treatments' physical aspects and GPR prediction. Improvements in data generalization and model performances can be associated with datasets having similar treatment conditions. This analysis might be implemented to evaluate the dataset quality and model consistency of further ML applications in radiotherapy. Dataset heterogeneities decrease ML model performance and reliability.

Interpretation the Influence of Hydrometeorological Variables on Soil Temperature Prediction Using the Potential of Deep Learning Model

The importance of soil temperature (ST) quantification can contribute to diverse ecological modelling processes as well as for agricultural activities. Over the literature, it was evident that soil supports more than 95% of living habitats and food production on earth, and this demand will increase to 500 years’ times in expected consumption in 2060. This paper aims to analyses the contrastive approach to predict the ST of a certain region with the help of different machine learning models, including Random Forest (RF), Support Vector, Neural Network (NN), Linear Regression (LR) and Long Short-Term Memory Network (LSTM). The study was utilized the hourly humidity, dew point, rainfall, solar radiation, and barometer readings for the formulation of the models. Various performance criteria were employed to evaluate the prediction skills of the models and the results depicted that the promising ability belong to LSTM despite the acceptable prediction accuracy achieved by other models. The modelling outcomes revealed that LSTM model attained the lowest root mean square error (RMSE = 3.3255) decreased the average prediction error by 6% with regards to NN (RMSE = 3.4796), SVM (RMSE = 3.5766), and RF (RMSE = 3.8128), and improved the prediction accuracy of LR by 15%. The model is in compliance with the latest machine learning industry standards and allows low-cost experimental performances on low powered edge computing devices.

Towards Managing Uncertain Geo-Information for Drilling Disasters Using Event Tracking Sensitivity Analysis.

In sub-surface drilling rigs, one key critical crisis is unwanted influx into the borehole as a result of increasing the influx rate while drilling deeper into a high-pressure gas formation. Although established risk assessments in drilling rigs provide a high degree of protection, uncertainty arises due to the behavior of the formation being drilled into, which may cause crucial situations at the rig. To overcome such uncertainties, real-time sensor measurements are used to predict, and thus prevent, such crises. In addition, new understandings of the effective events were derived from raw data. In order to avoid the computational overhead of input feature analysis that hinders time-critical prediction, EventTracker sensitivity analysis, an incremental method that can support dimensionality reduction, was applied to real-world data from 1600 features per each of the 4 wells as input and 6 time series per each of the 4 wells as output. The resulting significant input series were then introduced to two classification methods: Random Forest Classifier and Neural Networks. Performance of the EventTracker method was understood correlated with a conventional manual method that incorporated expert knowledge. More importantly, the outcome of a Neural Network Classifier was improved by reducing the number of inputs according to the results of the EventTracker feature selection. Most important of all, the generation of results of the EventTracker method took fractions of milliseconds that left plenty of time before the next bunch of data samples.

Prediction of compressive strength of geopolymer concrete using a hybrid ensemble of grey wolf optimized machine learning estimators

Geopolymer concrete (GPC) has the potential to replace conventional concrete. But, the mixed proportion of GPC poses several difficulties due to various contributing factors. The design of a reliable prediction model becomes challenging because of the non-linear relationship between the proportion of GPC mix and compressive strength (CS). This study implements a hybrid ensemble machine learning model (HEML) from grey wolf optimized conventional machine learning (CML) models to predict the CS of GPC. An experimental database of 1123 records was compiled from different published research work. The database was used to train and test three optimized CML models namely random forest regressor (RFR), Neural network (NN), and Multivariate regression spline (MARS). Utilizing a meta-learner XGBoost algorithm, the CML models' predicted outputs were trained to develop HEML. Statistical parameters notably R2, Adjusted R2, RMSE, MAE, MAPE, VAF, index, and the Shapiro-Wilk statistical test were used to compare the HEML and CML models' predicted outputs. The HEML model showed better performance in both the training and testing phase than the CML models. Two sensitivity analysis methods were adopted to analyze the impact of various parameters on the compressive strength of geopolymer concrete based on the model with the highest prediction accuracy.

Learning long-term motor timing/patterns on an orthogonal basis in random neural networks

The ability of the brain to generate complex spatiotemporal patterns with specific timings is essential for motor learning and temporal processing. An approach that can model this function, using the spontaneous activity of a random neural network (RNN), is associated with orbital instability. We propose a simple system that learns an arbitrary time series as the linear sum of stable trajectories produced by several small network modules. New finding in computer experiments is that the trajectories of the module outputs are orthogonal to each other. They created a dynamic orthogonal basis acquiring a high representational capacity, which enabled the system to learn the timing of extremely long intervals, such as tens of seconds for a millisecond computation unit, and also the complex time series of Lorenz attractors. This self-sustained system satisfies the stability and orthogonality requirements and thus provides a new neurocomputing framework and perspective for the neural mechanisms of motor learning.

Combinations of Scheimpflug tomography, ocular coherence tomography and air-puff tonometry improve the detection of keratoconus

PurposeTo determine whether combinations of devices with different measuring principles, supported by artificial intelligence (AI), can improve the diagnosis of keratoconus (KC). MethodsScheimpflug tomography, spectral-domain optical coherence tomography (SD-OCT), and air-puff tonometry were performed in all eyes. The most relevant machine-derived parameters to diagnose KC were determined using feature selection. The normal and forme fruste KC (FFKC) eyes were divided into training and validation datasets. The selected features from a single device or different combinations of devices were used to develop models based on random forest (RF) or neural networks (NN) trained to distinguish FFKC from normal eyes. The accuracy was determined using receiver operating characteristic (ROC) curves, area under the curve (AUC), sensitivity, and specificity. Results271 normal eyes, 84 FFKC eyes, 85 early KC eyes, and 159 advanced KC eyes were included. A total of 14 models were built. Air-puff tonometry had the highest AUC for detecting FFKC using a single device (AUC = 0.801). Among all two-device combinations, the highest AUC was accomplished using RF applied to selected features from SD-OCT and air-puff tonometry (AUC = 0.902), followed by the three-device combination with RF (AUC = 0.871) with the best accuracy. ConclusionExisting parameters can precisely diagnose early and advanced KC, but their diagnostic ability for FFKC could be optimized. Applying an AI algorithm to a combination of air-puff tonometry with Scheimpflug tomography or SD-OCT could improve FFKC diagnostic ability. The improvement in diagnostic ability by combining three devices is modest.

Development of machine learning models for the surveillance of colon surgical site infections

Conventional surgical site infection (SSI) surveillance is labour-intensive. We aimed to develop machine learning (ML) models for the surveillance of SSIs for colon surgery and to assess whether the ML could improve surveillance process efficiency. This study included cases who underwent colon surgery at a tertiary center between 2013 and 2014. Logistic regression and four ML algorithms including random forest (RF), gradient boosting (GB), and neural networks (NNs) with or without recursive feature elimination (RFE) were first trained on the entire cohort, and then re-trained on cases selected based on a previous rule-based algorithm. We assessed model performance based on the area under the curve (AUC), sensitivity, and positive predictive value (PPV). The estimated proportion of reduction in workload for chart review based on the ML models was evaluated and compared with the conventional method. At a sensitivity of 95%, the NN with RFE using 29 variables had the best performance with an AUC of 0.963 and PPV of 21.1%. When combining both the rule-based algorithm and ML algorithms, the NN with RFE using 19 variables had a higher PPV (28.9%) than with the ML algorithm alone, which could decrease the number of cases requiring chart review by 83.9% compared with the conventional method. We demonstrated that ML can improve the efficiency of SSI surveillance for colon surgery by decreasing the burden of chart review while providing high sensitivity. In particular, the hybrid approach of ML with a rule-based algorithm showed the best performance in terms of PPV.

Parsimonious physics-informed random projection neural networks for initial value problems of ODEs and index-1 DAEs.

We present a numerical method based on random projections with Gaussian kernels and physics-informed neural networks for the numerical solution of initial value problems (IVPs) of nonlinear stiff ordinary differential equations (ODEs) and index-1 differential algebraic equations (DAEs), which may also arise from spatial discretization of partial differential equations (PDEs). The internal weights are fixed to ones while the unknown weights between the hidden and output layer are computed with Newton's iterations using the Moore-Penrose pseudo-inverse for low to medium scale and sparse QR decomposition with L 2 regularization for medium- to large-scale systems. Building on previous works on random projections, we also prove its approximation accuracy. To deal with stiffness and sharp gradients, we propose an adaptive step-size scheme and address a continuation method for providing good initial guesses for Newton iterations. The "optimal" bounds of the uniform distribution from which the values of the shape parameters of the Gaussian kernels are sampled and the number of basis functions are "parsimoniously" chosen based on bias-variance trade-off decomposition. To assess the performance of the scheme in terms of both numerical approximation accuracy and computational cost, we used eight benchmark problems (three index-1 DAEs problems, and five stiff ODEs problems including the Hindmarsh-Rose neuronal model of chaotic dynamics and the Allen-Cahn phase-field PDE). The efficiency of the scheme was compared against two stiff ODEs/DAEs solvers, namely, ode15s and ode23t solvers of the MATLAB ODE suite as well as against deep learning as implemented in the DeepXDE library for scientific machine learning and physics-informed learning for the solution of the Lotka-Volterra ODEs included in the demos of the library. A software/toolbox in Matlab (that we call RanDiffNet) with demos is also provided.

QUANTITATIVE AND EXPLAINABLE ARTIFICIAL INTELLIGENCE (AI)-POWERED APPROACHES TO PREDICT ULCERATIVE COLITIS DISEASE ACTIVITY FROM HEMATOXYLIN AND EOSIN (H&E)-STAINED WHOLE SLIDE IMAGES (WSI)

Microscopic inflammation has been shown to be an important indicator of disease activity in ulcerative colitis (UC). However, manual histologic scoring is semi-quantitative and subject to interobserver variation, and AI-based solutions often lack interpretability. Here we report two distinct quantitative approaches to predict disease activity scores and histological remission using AI-powered digital pathology. Both the random forest classifier (RFC) and graph neural network (GNN) further provide explainability and biological insight by identifying histological features informing model predictions.

Adaption of Best Methodology by Comparing Different Algorithm to Detect Parkinson Disease

Abstract: Early and accurate prediction of Parkinson's disease (PD) can play a crucial role in improving its management and outcome. The aim of this study was to identify the best methodology for PD prediction. A comprehensive review of various prediction methods, including machine learning algorithms such as Support Vector Machines, Random Forest, and Neural Networks, was conducted. The performance of these methods was evaluated using metrics such as accuracy, sensitivity, and specificity. The findings revealed that a combination of Random Forest and Neural Networks emerged as the best approach for PD prediction, providing high accuracy and robust performance. This highlights the significance of selecting the appropriate prediction method for PD and the advantages of using a combination of algorithms for improved prediction.

A novel approach for surveying flowers as a proxy for bee pollinators using drone images

The abundance and diversity of plants and insects are important indicators of biodiversity, overall ecosystem health and agricultural production. Bees in particular are interesting indicators as they provide a key ecosystem service in many agricultural crops. Worldwide, habitat loss and fragmentation, agricultural intensification and climate change are important drivers of plant and bee decline. Monitoring of plants and bees is a crucial first step to safeguard their diversity and the services they provide but traditional in situ methods are time consuming and expensive. Remote sensing and Earth observation have the advantages that they can cover large areas and provides repeated, spatially continuous and standardized information. However, to date it has proven challenging to use these methods to assess small-scaled species-level biodiversity components with this approach. Here we surveyed bees and flowering plants using conventional field methods in 30 grasslands along a land-use intensity gradient in the Southeast of the Netherlands. We took RGB (true colored Red-Green–Blue) images using an Unmanned Aerial Vehicle (UAV) from the same fields and tested whether remote sensing can provide accurate assessments of flower cover and diversity and, by association, bee abundance and diversity. We explored the performance of different machine learning methods: Random Forest (RF), Neural Networks (NNET) and Support-Vector Machine (SVM). To evaluate the effect of the spatial resolution on the accuracy of the estimates, we tested all approaches using images at the original spatial resolution (∼ 0.5 cm) and re-sampled at 1 cm, 2 cm and 5 cm. We generally found significant relationships between UAV RGB derived estimates of flower cover and in situ estimates of flower cover and bee abundance and diversity. The highest resolution images generally resulted in the strongest relationships, with RF and NNET methods producing considerably better results than SVM methods (flower cover RF R2 = 0.8, NNET R2 = 0.79; bee abundance RF R2 = 0.65, NNET R2 = 0.54, bee species richness RF R2 = 0.62, NNET R2 = 0.52; bee species diversity RF R2 = 0.54, NNET R2 = 0.46). Our results suggest that methods based on the coupling of UAV imagery and machine learning methods can be developed into valuable tools for large-scale, standardized and cost-effective monitoring of flower cover and therefore of an important aspect of habitat quality for bees.

Computationally efficient Bayesian unit-level random neural network modelling of survey data under informative sampling for small area estimation

Abstract The topic of neural networks has seen a surge of interest in recent years. However, one of the main challenges with these approaches is quantification of uncertainty. The use of random weight models offer a potential solution. In addition to uncertainty quantification, these models are extremely computationally efficient as they do not require optimisation through stochastic gradient descent. We show how this approach can be used to account for informative sampling of survey data through the use of a pseudo-likelihood. We illustrate the effectiveness of this methodology through simulation and data application involving American National Election Studies data.