Search

Simple Summary: Breast cancer is a significant cause of female cancer-related deaths in the US. Checking how severe the cancer is helps in planning treatment. Modern AI methods are good at grading cancer, but they are not used much in hospitals yet. We developed and utilized ensemble deep learning algorithms for addressing the tasks of classifying (1) breast cancer subtype and (2) breast cancer invasiveness from whole slide image (WSI) histopathology slides. The ensemble models used were based on convolutional neural networks (CNNs) known for extracting distinctive features crucial for accurate classification. In this paper, we provide a comprehensive analysis of these models and the used methodology for breast cancer diagnosis tasks. Cancer diagnosis and classification are pivotal for effective patient management and treatment planning. In this study, a comprehensive approach is presented utilizing ensemble deep learning techniques to analyze breast cancer histopathology images. Our datasets were based on two widely employed datasets from different centers for two different tasks: BACH and BreakHis. Within the BACH dataset, a proposed ensemble strategy was employed, incorporating VGG16 and ResNet50 architectures to achieve precise classification of breast cancer histopathology images. Introducing a novel image patching technique to preprocess a high-resolution image facilitated a focused analysis of localized regions of interest. The annotated BACH dataset encompassed 400 WSIs across four distinct classes: Normal, Benign, In Situ Carcinoma, and Invasive Carcinoma. In addition, the proposed ensemble was used on the BreakHis dataset, utilizing VGG16, ResNet34, and ResNet50 models to classify microscopic images into eight distinct categories (four benign and four malignant). For both datasets, a five-fold cross-validation approach was employed for rigorous training and testing. Preliminary experimental results indicated a patch classification accuracy of 95.31% (for the BACH dataset) and WSI image classification accuracy of 98.43% (BreakHis). This research significantly contributes to ongoing endeavors in harnessing artificial intelligence to advance breast cancer diagnosis, potentially fostering improved patient outcomes and alleviating healthcare burdens. [ABSTRACT FROM AUTHOR]

Influenza affects millions of people every year. It causes a considerable amount of medical visits and hospitalisations as well as hundreds of thousands of deaths. Forecasting influenza prevalence with good accuracy can significantly help public health agencies to timely react to seasonal or novel strain epidemics. Although significant progress has been made, influenza forecasting remains a challenging modelling task. In this paper, we propose a methodological framework that improves over the state-of-the-art forecasting accuracy of influenza-like illness (ILI) rates in the United States. We achieve this by using Web search activity time series in conjunction with historical ILI rates as observations for training neural network (NN) architectures. The proposed models incorporate Bayesian layers to produce associated uncertainty intervals to their forecast estimates, positioning themselves as legitimate complementary solutions to more conventional approaches. The best performing NN, referred to as the iterative recurrent neural network (IRNN) architecture, reduces mean absolute error by 10.3% and improves skill by 17.1% on average in nowcasting and forecasting tasks across 4 consecutive flu seasons. Author summary: Modelling the prevalence of an infectious disease enables public health organisations to prepare for and minimise its impact. However, traditional disease indicators are often quite restrictive as they provide information with a significant delay. Recent research efforts have provided evidence of the value of alternative information sources such as Web search activity trends. Our work incorporates this information into influenza forecasting models to achieve state-of-the-art accuracy. In addition, the proposed Bayesian neural network architectures also provide associated uncertainty estimates for the forecasts, positioning our methodology as a practical complementary tool for disease surveillance and policy making. [ABSTRACT FROM AUTHOR]

Floods are significant global hazards that generally lead to loss of lives and billions worth of properties, especially in flood-prone regions. Various excellent artificial intelligence algorithms that are aimed at predicting streamflow to minimize the effects of floods are proposed by researchers. Most of these models depend on the assumption that both training and testing data set are similar and sufficient. However, in reality, many of these data set varies along with time and are insufficient in some new basins. Motivated by the success of transfer learning in natural language processing, image processing and time series forecasting. In this paper, we proposed two hybrid transfer learning models for streamflow forecasting. The proposed models, which integrate Gated Recurrent Unit (GRU) with transfer learning and integration of Long Short Term Memory (LSTM) with transfer learning, are compared to our reference model. The proposed coupled Transfer Learning with the GRU (TL+GRU) model outperforms the baseline models, i.e., the transfer learning model and the coupled Transfer Learning with LSTM model (TL+LSTM) for most of the basins when streamflow and precipitation data set from Model Parameter Estimation Experiment (MOPEX) basins in the United States of America is used. As a result, we can finally conclude that, with Artificial Neural Networks' (ANN) integration to transfer learning, more enhanced performance are obtained. [ABSTRACT FROM AUTHOR]

Annual turnover of home care workers represents a huge loss of revenue and is a key source of inefficiency in the home health care industry. In this article, we propose a data-driven approach to monitor employee churn and to capture the evolution of employee intent to leave. Unlike most papers in the literature, we use machine learning techniques to analyze over 2 million visits in the US, Canada, and Australia between 2016 and 2019. Results show that the gap between the number of hours worked and in the contract is the most important factor to predict employee intent to leave, which means an employee should be given as many hours as requested in the contract to improve retention. Secondary results show that having diverse shift lengths and continuity in services and patients seem to be associated with less turnover. [ABSTRACT FROM AUTHOR]

Water distribution network (WDN) leakage is estimated to account for more than 20% of potable water in the United States. Because existing leakage detection practices are mainly laborious or require excessive budget allocation, novel model-based techniques have become prevalent by employing smart meters. Specifically, incorporating machine learning and metaheuristic optimization methods into such model-based approaches to predict water pipeline leakage is one of the state-of-the-art approaches. This paper presents a predictive leakage detection approach by leveraging advanced metering infrastructure (AMI) and cyber-monitoring data. Primarily, the synchronous collection of consumption and monitoring data (i.e., flow and pressure) constitutes the proposed predictive tool in a combinative framework of genetic algorithms (GA) and particle swarm optimization (PSO). The time-consuming EPANET 2.2 simulation toolkit in MATLAB is circumvented by a set of trained artificial neural networks (ANNs). Sensitivity analyses are conducted to investigate the sensitivity of the proposed leakage prediction approach to multiple parameters. Ultimately, the number and location of the flow and pressure monitoring stations are optimized using maximum sensor availability thresholds. Mean absolute percentage error (MAPE) is employed to measure the model's accuracy. The proposed predictive tool offers insightful horizons on how considerable background leakage can be detected and prioritized without using labor-intensive, costly methods. [ABSTRACT FROM AUTHOR]

Technology forecasting is an important and critical issue that determines the starting point of planning and is considered as a management tool directly related to the future. In the previous research items, the development of renewable energy technologies was of concern. Moreover, due to the increasing need of countries to produce electricity and facing the lack of resources, this research focuses on forecasting photovoltaic technology. Accordingly, in this paper, for technological research in the field of solar energy, the patents extracted from one of the most famous renewable energy databases in the United States (US patent database) between 200 and 2020 were examined. Next, research gaps were analyzed by using the artificial neural network clustering method and also by analyzing covered and uncovered compounds. The results show that in the future, photovoltaic technology research will move towards the third generation of technology (organic materials) as well as focus on environmental parameters and their effects on the performance of photovoltaic systems. [ABSTRACT FROM AUTHOR]

Conventional statistical postprocessing techniques offer limited ability to improve the skills of probabilistic guidance for heavy precipitation. This paper introduces two artificial neural network (ANN)-based, geographically aware, and computationally efficient postprocessing schemes, namely, the ANN-multiclass (ANN-Mclass) and the ANN–censored, shifted gamma distribution (ANN-CSGD). Both schemes are implemented to postprocess Global Ensemble Forecast System (GEFS) forecasts to produce probabilistic quantitative precipitation forecasts (PQPFs) over the contiguous United States (CONUS) using a short (60 days), rolling training window. The performances of these schemes are assessed through a set of hindcast experiments, wherein postprocessed 24-h PQPFs from the two ANN schemes were compared against those produced using the benchmark quantile mapping algorithm for lead times ranging from 1 to 8 days. Outcomes of the hindcast experiments show that ANN schemes overall outperform the benchmark as well as the raw forecast over the CONUS in predicting probability of precipitation over a range of thresholds. The relative performance varies among geographic regions, with the two ANN schemes broadly improving upon quantile mapping over the central, south, and southeast, and slightly underperforming along the Pacific coast where skills of raw forecasts are the highest. Between the two schemes, the hybrid ANN-CSGD outperforms at higher rainfall thresholds (i.e., >50 mm day−1), though the outperformance comes at a slight expense of sharpness and spatial specificity. Collectively, these results confirm the ability of the ANN algorithms to produce skillful PQPFs with a limited training window and point to the prowess of the hybrid scheme for calibrating PQPFs for rare-to-extreme rainfall events. [ABSTRACT FROM AUTHOR]

In this paper we present a novel methodology, referenced as ORFEO (Ordinal classifier and Regressor Fusion for Estimating an Ordinal categorical target), to enhance the performance in ordinal classification problems for which the latent variable is observable. ORFEO is an artificial neural network model incorporating two outputs, one for ordinal classification, using the cumulative link model, and one for regression, using a linear model. Both outputs are simultaneously optimised considering a loss function that linearly combines both classification and regression losses. The main motivation behind developing the proposed approach is to enhance the performance of a standard ordinal classifier. This improvement is facilitated by considering the regression output, which allows the model to differentiate between patterns within the same category. The ORFEO model is applied to two problems in the field of marine and ocean engineering: short-term prediction of both significant wave height and flux of energy. Both problems are addressed considering four different coastal zones of the United States of America, using 13 datasets formed by buoys measurements and reanalysis data. A comprehensive comparison against 20 methodologies, including regression and nominal/ordinal classification approaches is performed, by using diverse nominal and ordinal performance metrics. Ranks achieved indicate that ORFEO outperforms all the compared methodologies in terms of all the performance measures, demonstrating the efficacy and robustness of the proposal. Finally, a statistical analysis is conducted, concluding that there are statistically significant differences across ordinal and nominal performance metrics in favour of the proposed ORFEO model. [ABSTRACT FROM AUTHOR]

Colon and rectal cancers are the most common kinds of cancer globally. Colon cancer is more prevalent in men than in women. Early detection increases the likelihood of survival, and treatment significantly increases the likelihood of eradicating the disease. The Surveillance, Epidemiology, and End Results (SEER) programme is an excellent source of domestic cancer statistics. SEER includes nearly 30% of the United States population, covering various races and geographic locations. The data are made public via the SEER website when a SEER limited-use data agreement form is submitted and approved. We investigate data from the SEER programme, specifically colon cancer statistics, in this study. Our objective is to create reliable colon cancer survival and conditional survival prediction algorithms. In this study, we have presented an overview of cancer diagnosis methods and the treatments used to cure cancer. This paper presents an analysis of prediction performance of multiple deep learning approaches. The performance of multiple deep learning models is thoroughly examined to discover which algorithm surpasses the others, followed by an investigation of the network's prediction accuracy. The simulation outcomes indicate that automated prediction models can predict colon cancer patient survival. Deep autoencoders displayed the best performance outcomes attaining 97% accuracy and 95% area under curve-receiver operating characteristic (AUC-ROC). [ABSTRACT FROM AUTHOR]

• This paper proposed a novel wind speed forecasting algorithm. • A new dataset division mechanism based on the moving window is used in this paper. • In order to improve the accuracy and stability simultaneously, the multi-objective particle swarm optimization is applied. • We compared the proposed MOPSO-Combined method with the combined model that published before in the paper. Accurate wind speed forecasting is important in power grid security, power system management, operation and market economics. However, most research has focused only on improving either accuracy or stability, with few studies addressing the two issues, simultaneously. Therefore, we proposed a novel combined model based on multi-objective particle swarm optimization, which is applied to optimize the key parameters of the echo state network. Most combined wind speed forecasting methods just use the combination theory to combine individual methods, this paper uses echo state network to combine the intermediate wind speed forecasting results of three artificial neural networks. Moreover, a new dataset division mechanism based on the moving window is applied in this paper. Firstly, the length of the input data is changed from 5 to 15 for 1-step, 2-step and 3-step wind speed forecasting, after that, the optimal length of the input vector can be got. And then we apply this optimal length of the input vector to another dataset for further verifying the proposed method. In order to verify the forecasting effectiveness of the proposed forecasting model, the 80/min wind speed data of M2 tower of the National Wind Power Technology Center of the United States were taken as an example. The experimental results indicate that the proposed algorithm is superior to the other ten comparative models in prediction accuracy and stability, and it also performs better than the combined model that we have proposed before. [ABSTRACT FROM AUTHOR]

The prediction of wave height and flux of energy is essential for most ocean engineering applications. To simultaneously predict both wave parameters, this paper presents a novel approach using short-term time prediction horizons (6h and 12h). Specifically, the methodology proposed presents a twofold simultaneity: 1) both parameters are predicted by a single model, applying the multi-task learning paradigm, and 2) the prediction tasks are tackled for several neighbouring ocean buoys with such single model by the development of a zonal strategy. Multi-Task Evolutionary Artificial Neural Network (MTEANN) models are applied to two different zones located in the United States, considering measurements collected by three buoys in each zone. Zonal MTEANN models have been compared in a two-phased procedure: 1) against the three individual MTEANN models specifically trained for each buoy of the zone, and 2) against some state-of-the-art regression techniques. Results achieved show that the proposed zonal methodology obtains not only better performance than the individual MTEANN models, but it also requires a lower number of connections. Besides, the zonal MTEANN methodology outperforms state-of-the-art regression techniques. Hence, the proposed approach results in an excellent method for predicting both significant wave height and flux of energy at short-term prediction time horizons. [ABSTRACT FROM AUTHOR]

Summary: Many electricity consumers around the world are billed electric peak load charges. In the United States, these demand charges can contribute up to 70% of a consumer's electricity bill. Research has shown that the financial impact of these charges to the consumers can be reduced by acting on the intelligence provided by peak electric load day (PELD) forecasts. Unfortunately, published research detailing PELDs forecasting methodologies for facilities adopting behind the meter renewable electricity generation (BTMREG) is non‐existent. This paper contributes: (a) a PELD forecasting methodology applicable to consumers with BTMREG. A case study demonstrates the methodology's ability to achieve 93% of the potential savings in kW, translating to demand charge reductions of approximately US$ 142 129 per year. (b) The first side‐by‐side performance comparisons of electric load and PELD forecasting models for scenarios with and without BTMREG. (c) The first PELD forecasting model savings comparison for scenarios with and without BTMREG. Experimental results suggest that counterintuitively, BTMREG adoption can translate into higher peak load savings for electricity consumers. [ABSTRACT FROM AUTHOR]

In this paper, we propose bag of adversarial features (BAFs) for identifying mild traumatic brain injury (MTBI) patients from their diffusion magnetic resonance images (MRIs) (obtained within one month of injury) by incorporating unsupervised feature learning techniques. MTBI is a growing public health problem with an estimated incidence of over 1.7 million people annually in USA. Diagnosis is based on clinical history and symptoms, and accurate, concrete measures of injury are lacking. Unlike most of the previous works, which use hand-crafted features extracted from different parts of brain for MTBI classification, we employ feature learning algorithms to learn more discriminative representation for this task. A major challenge in this field thus far is the relatively small number of subjects available for training. This makes it difficult to use an end-to-end convolutional neural network to directly classify a subject from MRIs. To overcome this challenge, we first apply an adversarial auto-encoder (with convolutional structure) to learn patch-level features, from overlapping image patches extracted from different brain regions. We then aggregate these features through a bag-of-words approach. We perform an extensive experimental study on a dataset of 227 subjects (including 109 MTBI patients, and 118 age and sex-matched healthy controls) and compare the bag-of-deep-features with several previous approaches. Our experimental results show that the BAF significantly outperforms earlier works relying on the mean values of MR metrics in selected brain regions. [ABSTRACT FROM AUTHOR]

Porosity is one of the most important parameters to assess in-place oil or gas in reservoirs, and to evaluate recovery from enhanced production operations. Since it is relatively well-established to determine porosity using different laboratory and field methods, its value is usually determined at many locations across a reservoir as part of the common practice to capture reservoir heterogeneity and the variability in values. This suite of measurements and the distribution of values are most valuable for probabilistic reservoir assessments, and for spatial modeling if the exact data locations are known. Despite the importance of individual measurements to set the range of values for probabilistic studies, it is not always possible to access these data due to confidentiality. In most cases, commercial or publicly available databases that assessments may rely on usually report only mean values of porosity, like any other reservoir data, or they may not report a value at all. This makes both quantifying the mean value and the uncertainty around it difficult for probabilistic assessments. In this study, the TORIS (Tertiary Oil Recovery Information System) dataset of the National Petroleum Council and the U.S. Department of Energy was used to model porosity and the uncertainty around predicted values. TORIS is an integrated dataset of production data, reservoir properties, and project databases of crude oil reservoirs in the United States. The model presented in the paper was based on ANCOVA (Analysis of Co-Variance) of data from 1038 reservoirs from the TORIS dataset for porosity prediction, validation and testing for quantitative and qualitative parameters that may be readily available in most cases, and to estimate uncertainty around the mean values. This model also explored association of porosity values to different parameters, and to different depositional systems and diagenetic overprint conditions. Furthermore, an ANN (Artificial Neural Network) model was created to compare the predicted values of both models. Results showed that the ANN model was able to represent more of the variability, however it lacked the insights that might be gained from the ANCOVA model. • TORIS dataset was used to model porosity and its uncertainty around the mean values using ANCOVA. • esides quantitative variables, qualitative variables were used. • A regression equation was generated, validated and tested using additional data. • The model can be used to predict porosity when data is scarce. • The model can also be used to generate prediction interval around the mean values for probabilistic assessments. [ABSTRACT FROM AUTHOR]

Malignancy is the tendency of a medical condition to progressively become fatal and generally refers to the presence of cancerous cells that can spread, invade, and destroy tissues. In the United States, a new cancer diagnosis is given every 30 s, and the rates of diagnosis are not indicating any slowdown. The number of cancer cases is projected to increase by 35% over the next decade. Increasing cancer cases have attracted researchers to find new approaches for their identification and characterization. Skin diseases are a significant healthcare burden, affecting millions of people globally. The early detection and diagnosis of skin diseases are essential to prevent adverse health consequences, including the development of malignant skin tumors. Dermoscopy-based diagnosis of skin diseases is a time-consuming process, particularly in resource-limited healthcare settings. Recent advancements in computer vision algorithms have led to the development of several smartphone-based applications for early skin disease prognosis. In this paper, we have utilized a set of 16 distinctive convolutional neural network models, based on deep learning, to process over 45,000 high-quality images from the HAM10000 dataset, encompassing seven categories of skin diseases. Our models were able to effectively identify and classify skin diseases using lesion images as input. Dermoscopy-based diagnosis of skin diseases is a time-consuming process, particularly in resource-limited healthcare settings. Our proposed application for early skin disease detection aims to prevent unintended health consequences, which can result in wastage of operating hours, resources, and, in some cases, compromised patient health. We have benchmarked our models against one another to evaluate their effectiveness and efficiency, using performance measures such as accuracy, precision, sensitivity, specificity, F-score, and G-Mean. Our results demonstrate that most deployed models achieve an accuracy of up to 99%. [ABSTRACT FROM AUTHOR]

BACKGROUND: The assessment of bone age and skeletal maturity and its comparison to chronological age is an important task in the medical environment for the diagnosis of pediatric endocrinology, orthodontics and orthopedic disorders, and legal environment in what concerns if an individual is a minor or not when there is a lack of documents. Being a time-consuming activity that can be prone to inter- and intra-rater variability, the use of methods which can automate it, like Machine Learning techniques, is of value. OBJECTIVE: The goal of this paper is to present the state of the art evidence, trends and gaps in the research related to bone age assessment studies that make use of Machine Learning techniques. METHOD: A systematic literature review was carried out, starting with the writing of the protocol, followed by searches on three databases: Pubmed, Scopus and Web of Science to identify the relevant evidence related to bone age assessment using Machine Learning techniques. One round of backward snowballing was performed to find additional studies. A quality assessment was performed on the selected studies to check for bias and low quality studies, which were removed. Data was extracted from the included studies to build summary tables. Lastly, a meta-analysis was performed on the performances of the selected studies. RESULTS: 26 studies constituted the final set of included studies. Most of them proposed automatic systems for bone age assessment and investigated methods for bone age assessment based on hand and wrist radiographs. The samples used in the studies were mostly comprehensive or bordered the age of 18, and the data origin was in most of cases from United States and West Europe. Few studies explored ethnic differences. CONCLUSIONS: There is a clear focus of the research on bone age assessment methods based on radiographs whilst other types of medical imaging without radiation exposure (e.g. magnetic resonance imaging) are not much explored in the literature. Also, socioeconomic and other aspects that could influence in bone age were not addressed in the literature. Finally, studies that make use of more than one region of interest for bone age assessment are scarce. [ABSTRACT FROM AUTHOR]

Synthetic Aperture Radar (SAR) remote sensing has been widely used for flood mapping and monitoring. Nevertheless, flood detection in urban areas still proves to be particularly challenging by using SAR. In this paper, we assess the roles of SAR intensity and interferometric coherence in urban flood detection using multi-temporal TerraSAR-X data. We further introduce an active self-learning convolution neural network (A-SL CNN) framework to alleviate the effect of a limited annotated training dataset. The proposed framework selects informative unlabeled samples based on a temporal-ensembling CNN model. These samples are subsequently pseudo-labeled by a multi-scale spatial filter. Consistency regularization is introduced to penalize incorrect labels caused by pseudo-labeling. We show results for a case study that is centered on flooded areas in Houston, USA, during hurricane Harvey in August 2017. Our experiments show that multi-temporal intensity (pre- and co-event) plays the most important role in urban flood detection. Adding multi-temporal coherence can increase the reliability of the inundation map considerably. Meanwhile, encouraging results are achieved by the proposed A-SL CNN framework: the к statistic is improved from 0.614 to 0.686 in comparison to its supervised counterpart. [ABSTRACT FROM AUTHOR]

• Coupling of ANN waste forecasts with GIS collection route optimization. • GIS routes with minimum travel distances are sensitive to waste composition. • Results of 36 scenarios yield changes in travel distance of up to 19.9% • Dual compartment trucks reduce travel distance but increase waste collection time. Combining an artificial neural network (ANN) waste prediction model with a geographic information system (GIS) waste collection route optimization, the paper shows how the compositional features of waste materials affect the optimized truck route time, distance, and air emissions. Using data from Austin, Texas, USA, a nonlinear autoregressive ANN model is used to predict the waste generation rate of the recycling and garbage streams for the year 2023 in four sub-areas of the city. This ANN model resulted in mean absolute percentage errors ranging from 10.92% to 16.51%. Modified compositions of the recycling and garbage streams are then used as inputs, along with the year 2023 generation rates, to create 6 modified and 3 non-modified scenarios that reflect possible future changes in waste composition. These waste stream scenarios are then used as input parameters to determine optimal waste collection routes with minimal travel distance in each of the four sub-areas using the GIS vehicle routing problem network analysis tool. Results of these 36 scenarios yield changes in travel distance of up to 19.9%, when compared to the non-modified composition. Further, dual compartment trucks were compared to single compartment trucks and found to save between 10.3 and 16.0% in travel distance and slightly reduce emissions but had a 15.7–19.8% increase in collection time. Results suggest temporal changes in waste composition and characteristics are important in GIS route optimization studies. [ABSTRACT FROM AUTHOR]

Overall performance of energy infrastructure in the United States has been assessed as D+. More than 65% of America's energy is transported through the oil and gas pipelines, which have experienced more than 10,000 failures during the last three decades. There is a critical need for a failure prediction tool that can forecast the consequences of the hazardous failures. Failure of gas pipelines has become the subject of interest for some studies in the past. Previous studies mainly focused on physical models that need inspection data or developed subjective models. This paper aims at developing a model to forecast the consequences of the potential failures of such pipes using the historical data of the U.S. gas pipes network. The model applies a neurofuzzy technique in order to recognize the existing pattern among the input and output variables. It estimates the financial consequences of various failure scenarios for specific pipes in terms of size and specified minimum yield strength. For this purpose, a bowtie model is developed, and all possible scenarios of failure are identified. Various combinations of the identified factors and different number and types of membership functions, are applied in order to optimize the model's efficiency. The developed model is validated with an approximate accuracy of 80%. This study assists practitioners and academics who are working on the risk assessment of gas pipelines to plan for their lifecycle inspection. [ABSTRACT FROM AUTHOR]

• More accurate and explanatory model is proposed for natural gas storage deliverability. • A hybrid Sparse Autoencoder and Random Forest model shows on average of 5.7% increase in prediction accuracy. • The first use of machine learning and explanatory methods (XAI) in deliverability prediction for natural gas storage. Numerous subsurface factors, including geology and fluid properties, can affect the connectivity of the storage spaces in depleted reservoirs; hence, fluid flow simulations become more complicated, and predicting their deliverability remains challenging. This paper applies Machine Learning (ML) techniques to predict the deliverability of underground natural gas storage (UNGS) in depleted reservoirs. First, three baseline models were developed based on Support Vector Regression (SVR), Artificial Neural Network (ANN), and Random Forest (RF) algorithms. To improve the accuracy of the RF model as the best-performing baseline model, a unified framework, referred to as SARF, was developed. SARF combines the capabilities of Sparse Autoencoder (SA) and that of Random Forest (RF). To achieve this, the internal representations of the SA, which constitute extracted features of the input variables, are used in RF to develop the proposed SARF framework. The predictive capabilities of the baseline models and the proposed SARF model were validated using 3744 real-world storage data samples of 52 active storage reservoirs in the United States. The experimental result of this study shows that the proposed SARF model achieved an average 5.7% increase in accuracy on four separate data partitions over the baseline RF model. Furthermore, a set of eXplainable Artificial Intelligence (XAI) methods were developed to provide an intuitive explanation of which factors influence the deliverability of reservoir storage. The visualizations developed using the XAI method provide an easy-to-understand interpretation of how the SARF model predicted the deliverability values for separate reservoirs. [ABSTRACT FROM AUTHOR]

This paper proposes a collection of novel deliverability prediction models for underground natural gas storage (UNGS) in salt caverns based on machine learning algorithms. Considering that the natural gas supply chain is characterized by imbalances between demand and supply on a timely basis, effective and fast models for predicting the deliverability of UNGS would not only be a valuable tool to various stakeholders but also, of great benefit in competitive natural gas marketplace. In this paper, a first step in applying machine learning algorithms to predict the deliverability of UNGS in salt caverns is proposed. To achieve this, the capability of three machine learning algorithms namely, artificial neural network (ANN), support vector machine (SVM), and Random Forest (RF) are examined. The predictive capabilities of these methods were investigated using different monthly field storage data samples for different years with varied data samples of 36 active UNGS in salt caverns in the United States. Experimental results reveal that the machine learning models developed in this study can serve as suitable tools for predicting the deliverability of UNGS in salt caverns with different performances. Overall result shows that RF model exhibits better prediction performance with varied data partitions over ANN and SVM models. • Deliverability prediction for underground natural gas storage in salt caverns. • Artificial Neural Network, Random Forest, and Support Vector Machines are examined. • Random Forest exhibits better prediction performance on different data samples. • Results can facilitate decisions for narrowing the gap between demand and supply. • Results can guide decisions for future design specification of gas storage. [ABSTRACT FROM AUTHOR]

Bank failure prediction is an important study for regulators in the banking industry because the failure of a bank leads to devastating consequences. If bank failures are correctly predicted, early warnings can be sent to the responsible authorities for precaution purposes. Therefore, a reliable bank failure prediction or early warning system is invaluable to avoid adverse repercussion effects on other banks and to prevent drastic confidence losses in the society. In this paper, we propose a novel self-organizing neural fuzzy inference system, which functions as an early warning system of bank failures. The system performs accurately based on the auto-generated fuzzy inference rule base. More importantly, the simplified rule base possesses a high level of interpretability, which makes it much easier for human users to comprehend. Three sets of experiments are conducted on a publicly available database, which consists of 3635 United States banks observed over a 21-year period. The experimental results of our proposed model are encouraging in terms of both accuracy and interpretability when benchmarked against other prediction models. [ABSTRACT FROM AUTHOR]

Effective energy planning and governmental decision-making policies heavily rely on accurate forecast of energy demand. This paper discusses and compares five different forecasting techniques to model energy demand in the United States using economic and demographic factors. Two artificial neural network (ANN) models, two regression analysis models, and one autoregressive integrated moving average (ARIMA) model are developed based on the historical data from 1950 to 2013. While ANN model I and regression model I use gross domestic product (GDP), gross national product (GNP), and per capita personal income as independent input factors, ANN model 2 and regression model 2 employ GDP, GNP, and population (POP) as the predictive factors. The forecasted values resulted from these models are compared with the forecast made by the US. Energy Information Administration (EIA) for the period of 2014-2019. The forecasted results of ANN models and regression model 1 are dose to those of the US. El A; however, the results of regression model 2 and ARIMA model are significantly different from the forecast made by the US. El A. Finally, a comparison of the forecasted values resulted from three efficient models showed that the energy demand would vary between 95.51 and 100.08 quadrillion British thermal unit (btu) for the period of 2014-2019. In addition, we have discussed the possibility of self-sufficiency of the United States in terms of energy generation based on the information of current available technologies nationwide. [ABSTRACT FROM AUTHOR]

In the United States, the industrial sector is the driving engine of economic development, and energy consumption in this sector may be considered as the fuel for this engine. In order to keep this sector sustainable (diverse and productive over the time), energy planning should be carried out comprehensively and precisely. This paper describes the development of two types of numerical energy models which are able to predict the United States' future industrial energy-demand. One model uses an ANN (artificial neural network) technique, and the other model uses a MLR (multiple linear regression) technique. Various independent variables (GDP, price of energy carriers) are tested. The future industrial energy demand can then be forecasted based on a defined scenario. The ANN model anticipates a 16% increase in energy demand from 2012 by 2030. In this forecast, the model assumes that the effective independent parameters remain constant during this period and only GDP grows with a second-order polynomial trend. The forecast result, which shows consistency with published predictions, may be considered as an indication of the need for development of new and low-cost energy sources. This study suggests that the ANN technique is a reliable and powerful technique which can effectively perform input/output mapping. In order to validate the performance of the models, the results of the ANN model is compared to the projections from the Energy Information Administration of the U.S. Department of Energy. [ABSTRACT FROM AUTHOR]

Liu, C., 2019. Attribution of marine crop production in the United States of America based on artificial neural network learning. In: Li, L.; Wan, X.; and Huang, X. (eds.), Recent Developments in Practices and Research on Coastal Regions: Transportation, Environment and Economy. Journal of Coastal Research Journal of Coastal Research, Special Issue No. 98, pp. 277–283. Coconut Creek (Florida), ISSN 0749-0208. Marine crops play an important role in the food chain and in agriculture. Marine crop products are used as livestock feed and to meet people's food requirements. This paper examines what factors cause differences in marine crop production and attribution to marine crop yield. The paper builds a marine crop production model based on an artificial neural network. Then it analyzes the sensitivity of all factors to understand their contributions to marine crop production. [ABSTRACT FROM AUTHOR]

Heavy-duty trucks contribute approximately 20% of fuel consumption in the United States of America (USA). The fuel economy of heavy-duty vehicles (HDV) is affected by several real-world parameters like road parameters, driver behavior, weather conditions, and vehicle parameters, etc. Although modern vehicles comply with emissions regulations, potential malfunction of the engine, regular wear and tear, or other factors could affect vehicle performance. Predicting fuel consumption per trip based on dynamic on-road data can help the automotive industry to reduce the cost and time for on-road testing. Data modeling can easily help to diagnose the reason behind fuel consumption with a knowledge of input parameters. In this paper, an artificial neural network (ANN) was implemented to model fuel consumption in modern heavy-duty trucks for predicting the total and instantaneous fuel consumption of a trip based on very few key parameters, such as engine load (%), engine speed (rpm), and vehicle speed (km/h). Instantaneous fuel consumption data can help to predict patterns in fuel consumption for optimized fleet operations. In this work, the data used for modeling was collected at a frequency of 1Hz during on-road testing of modern heavy-duty vehicles (HDV) at the West Virginia University Center for Alternative Fuels Engines and Emissions (WVU CAFEE) using the portable emissions monitoring system (PEMS). The performance of the artificial neural network was evaluated using mean absolute error (MAE) and root mean square error (RMSE). The model was further evaluated with data collected from a vehicle on-road trip. The study shows that artificial neural networks performed slightly better than other machine learning techniques such as linear regression (LR), and random forest (RF), with high R-squared (R2) and lower root mean square error. [ABSTRACT FROM AUTHOR]

Malignant melanoma accounts for about 1–3% of all malignancies in the West, especially in the United States. More than 9000 people die each year. In general, it is difficult to characterize a skin lesion from a photograph. In this paper, we propose a deep learning-based computer-aided diagnostic algorithm for the classification of malignant melanoma and benign skin tumors from RGB channel skin images. The proposed deep learning model constitutes a tumor lesion segmentation model and a classification model of malignant melanoma. First, U-Net was used to classify skin lesions in dermoscopy images. We implement an algorithm to classify malignant melanoma and benign tumors using skin lesion images and expert labeling results from convolutional neural networks. The U-Net model achieved a dice similarity coefficient of 81.1% compared to the expert labeling results. The classification accuracy of malignant melanoma reached 80.06%. As a result, the proposed AI algorithm is expected to be utilized as a computer-aided diagnostic algorithm to help early detection of malignant melanoma. [ABSTRACT FROM AUTHOR]

• Temporal and spatial features are used for wind speed forecasting. • A novel probabilistic deep learning model is proposed for wind speed forecasting. • The proposed model shows the best performance compared with other models. • Temporal and spatial uncertainty of the wind speed are estimated based on the model. Reliable and accurate probabilistic forecasting of wind speed is of vital importance for the utilization of wind energy and operation of power systems. In this paper, a probabilistic spatiotemporal deep learning model for wind speed forecasting is proposed. The underlying wind turbines are embedded into a grid space, which fully expresses the spatiotemporal variation process of the airflow. Thus, advanced image recognition methods can be employed to solve the spatiotemporal wind speed forecasting problem. The proposed model is based on a spatial–temporal neural network (STNN) and variational Bayesian inference. The proposed STNN combines the convolutional GRU model and 3D Convolutional Neural Network and uses an encoding-forecasting structure to generate the spatiotemporal predictions. Variational Bayesian inference is employed to obtain the approximated posterior parameter distribution of the model and determine the probability of the prediction. The proposed model is applied in two real-world case studies in United States. The experimental results demonstrate that the proposed model significantly outperforms other models in both forecast skill and forecast reliability. The uncertainty estimation is also shown and it demonstrates that the proposed model is able to provide effective uncertainty estimation in both the time level and space level. [ABSTRACT FROM AUTHOR]

• Successful disaster management requires data describing disaster impact and extent of damage. • Aerial data collection is resource-intensive and requires offline post-processing. • We introduce convolutional neural network (CNN) models for ground object detection. • The models recognize building roofs, cars, vegetation, debris, and flooded areas. • We achieve 80.69% and 74.48% mAP for high and low altitude footage, respectively. Accurate and timely access to data describing disaster impact and extent of damage is key to successful disaster management (a process that includes prevention, mitigation, preparedness, response, and recovery). Airborne data acquisition using helicopter and unmanned aerial vehicle (UAV) helps obtain a bird's-eye view of disaster-affected areas. However, a major challenge to this approach is robustly processing a large amount of data to identify and map objects of interest on the ground in real-time. The current process is resource-intensive (must be carried out manually) and requires offline computing (through post-processing of aerial videos). This research introduces and evaluates a series of convolutional neural network (CNN) models for ground object detection from aerial views of disaster's aftermath. These models are capable of recognizing critical ground assets including building roofs (both damaged and undamaged), vehicles, vegetation, debris, and flooded areas. The CNN models are trained on an in-house aerial video dataset (named Volan2018) that is created using web mining techniques. Volan2018 contains eight annotated aerial videos (65,580 frames) collected by drone or helicopter from eight different locations in various hurricanes that struck the United States in 2017–2018. Eight CNN models based on You-Only-Look-Once (YOLO) algorithm are trained by transfer learning, i.e., pre-trained on the COCO/VOC dataset and re-trained on Volan2018 dataset, and achieve 80.69% mAP for high altitude (helicopter footage) and 74.48% for low altitude (drone footage), respectively. This paper also presents a thorough investigation of the effect of camera altitude, data balance, and pre-trained weights on model performance, and finds that models trained and tested on videos taken from similar altitude outperform those trained and tested on videos taken from different altitudes. Moreover, the CNN model pre-trained on the VOC dataset and re-trained on balanced drone video yields the best result in significantly shorter training time. [ABSTRACT FROM AUTHOR]

Since wildfires have occurred frequently in recent years, accurate burned area mapping is required for wildfire severity assessment and burned land reconstruction. Satellite remote sensing is an effective technology that can provide valuable information for wildfire assessment. However, the common approaches based on using a single satellite image to promptly detect the burned areas have low accuracy and limited applicability. This paper develops a new burned area mapping method that surpasses the detection accuracy of previous methods, while still using a single Moderate Resolution Imaging Spectroradiometer (MODIS) sensor image. The key innovation is integrating optimal spectral indices and a neural network algorithm. We used the traditional empirical formula method, multi-threshold method and visual interpretation method to extract the sample sets of five typical types (burned area, vegetation, cloud, bare soil, and cloud shadow) from the MODIS data of several wildfires in the American states of Nevada, Washington and California in 2016. Afterward, the separability index M was adopted to assess the capacity of seven spectral bands and 13 spectral indices to distinguish the burned area from four unburned land cover types. Based on the separability analysis between the burned area and unburned areas, the spectral indices with an M value higher than 1.0 were employed to generate the training sample sets that were assessed to have an overall accuracy of 98.68% and Kappa coefficient of 97.46%. Finally, we utilized a back-propagation neural network (BPNN) to learn the spectral differences of different types from the training sample sets and obtain the output burned area map. The proposed method was applied to three wildfire cases in the American states of Idaho, Nevada and Oregon in 2017. A comparison of detection results between the new MODIS-based burned area map and the reference burned area map compiled from Landsat-8 Operational Land Imager (OLI) data indicates that the proposed method can effectively exploit the spectral characteristics of various land cover types. Also, this new method can achieve higher accuracy with the reduction of commission error (CE, >10%) and omission error (OE, >6%) compared to the traditional empirical formula method. The new burned area mapping method could help managers and the public perform more effective wildfire assessments and emergency management. [ABSTRACT FROM AUTHOR]