Search

Aiming at the problem of long soil–water characteristic curve (SWCC) testing times and the difficulty of prediction accuracy in complex environments, this paper establishes a SWCC prediction model based on a neural network machine learning algorithm which can take into account the influence of multiple factors such as temperature, deformation, and salinity. The input layer of the model can reflect the physical properties of the soil and the influence of the external environment, while the suction is taken as an input variable, which in turn can directly obtain the water content under the corresponding conditions. The predictive ability of the model is verified by comparing and analyzing the predicted results of the SWCC under different temperature, void ratio, and salinity conditions with the experimental results. The research in this paper provides a new method for predicting the SWCC considering multiple factors, and the prediction accuracy of the model is related to the amount of experimental data. [ABSTRACT FROM AUTHOR]

Sindh province of Pakistan has a long history of severe droughts. Several large scale climate drivers (LSCD) are known for their effect on precipitation worldwide but studies in the Sindh region are missing; wide variety of LSCDs and lagged associative information. This study aimed to identify the significant LSCDs in Sindh province of Pakistan and improve the forecast skill of monthly precipitation by employing the principal component analysis (PCA), artificial neural network (ANN), Bayesian regularization neural network (BRNN), and multiple regression analysis (MRA), while considering the 12 months lagged LSCDs such as Nino-1+2, Nino-3, Nino-3.4, Nino-4, Quasi-Biennial Oscillation (QBO) at 30 and 50hPa (QBOI and QBOII), sea surface temperature (SST), 2m air temperature (T2M), 500 hPa and 850 hPa geopotential heights (H500 and H850), surface and 500 hPa zonal velocity (SU and U500), latent and sensible heat fluxes over land (LHFOL and SHFOL), and surface specific humidity (SSH). Global Land Data Assimilation System (GLDAS), Tropical Rainfall Measuring Mission (TRMM), Modern-Era Retrospective Analysis for Research and Application (MERRA-2), NOAA, Freie University Berlin, and Hadley Centre Sea Ice and Sea Surface Temperature (HadISST) datasets were used. Results manifested that significant LSCDs with 99% confidence level were SU, U500, T2M, SST, SHFOL, LHFOL, SSH, and H850. During test period, compared with MR models of 0.39 to 0.64 and principal components of 0.31 to 0.57, the ANN and BRNN models had better predictive skills with correlation coefficients of 0.57 to 0.83 and 0.52 to 0.76, respectively. It can be concluded that the ANN and BRNN models enable us to predict monthly precipitation in Sindh region with lagged LSCDs. [ABSTRACT FROM AUTHOR]

The present study utilized response surface methodology (RSM) and Bayesian neural network (BNN) to predict the characteristics of a diesel engine powered by a blend of biodiesel and diesel fuel. The biodiesel was produced from waste cooking oil using a biocatalyst synthesized from vegetable waste through the wet impregnation technique. A multilevel central composite design was utilized to predict engine characteristics, including brake thermal efficiency (BTE), nitric oxide (NO), unburned hydrocarbons (UBHC), smoke emissions, heat release rate (HRR), and cylinder peak pressure (CGPP). BNN and the logistic–sigmoid activation function were used to train the experimental data in the artificial neural network (ANN) model, and the errors and correlations of the predicted models were calculated. The study revealed that the biocatalyst was capable of producing a maximum yield of 93% at 55 °C under specific reaction conditions, namely a reaction time of 120 min, a stirrer speed of 900 rpm, a catalyst loading of 7 wt.%, and a molar ratio of 1:9. Further, the ANN model was found to exhibit comparably lower prediction errors (0.001–0.0024), lower MAPE errors (3.14–4.6%), and a strong correlation (0.984–0.998) compared to the RSM model. B100-80%-20% was discovered to be the best formulation for emission property, while B100-90%-10% was the best mix for engine performance and combustion at 100% load. In conclusion, this study found that utilizing the synthesized biocatalyst led to attaining a maximum biodiesel yield. Furthermore, the study recommends using ANN and RSM techniques for accurately predicting the characteristics of a diesel engine. [ABSTRACT FROM AUTHOR]

Overfitting often occurs in neural network training, and neural networks with higher generalization ability are less prone to this phenomenon. Aiming at the problem that the generalization ability of photovoltaic (PV) power prediction model is insufficient, a PV power time-sharing prediction (TSP) model combining variational mode decomposition (VMD) and Bayesian regularization neural network (BRNN) is proposed. Firstly, the meteorological sequences related to the output power are selected by mutual information (MI) analysis. Secondly, VMD processing is performed on the filtered sequences, which is aimed at reducing the non-stationarity of the data; then, normalized cross-correlation (NCC) and signal-to-noise ratio (SNR) between the components obtained by signal decomposition and the original data are calculated, after which the key influencing factors are screened out to eliminate the correlation and redundancy of the data. Finally, the filtered meteorological sequences are divided into two datasets based on whether the irradiance of the day is zero or not. Meanwhile, the predictions are performed using BRNN for each of the two datasets. Then, the results are reordered in chronological order, and the prediction of PV power is realized conclusively. It was experimentally verified that the mean absolute value error ( M A E ) of the method proposed in this paper is 0.1281 , which is reduced by 40.28 % compared with the back propagation neural network (BPNN) model on the same dataset, the mean squared error ( M S E ) is 0.0962 , and the coefficient of determination ( R 2 ) is 0.9907 . Other error indicators also confirm that VMD is of much significance and TSP is contributive. [ABSTRACT FROM AUTHOR]

The Balochistan province of Pakistan is mostly affected by severe drought events due to little amount of precipitation. “Several Large Scale Climate Drivers (LSCDs) are known for their effects on precipitation worldwide but studies in the region are missing; a wide variety of LSCDs and lagged associative information”.The current study aimed to identify the significant LSCDs in the Balochistan province of Pakistan and improve the prediction skill of monthly precipitation by employing the Principal Component Analysis, Artificial Neural Network (ANN), Bayesian Regularization Neural Network (BRNN), and Multiple Regression (MR) Analysis using the 12 months lagged LSCDs such as Nino-1+2, Nino-3, Nino-3.4, Nino-4, QBO at 30 and 50hpa (QBOI and QBOII), Sea Surface Temperature (SST), 2m air temperature (T2M), 500hpa and 850hpa geopotential heights (H500 and H850), 500hpa zonal velocity (U500), 500hpa and 850hpa meridional velocity (V500 and V850), Latent and Sensible Heat Fluxes Over Land (LHFOL and SHFOL), and Surface Specific Humidity (SSH). To collect the data, Global Land Data Assimilation System, Tropical Rainfall Measuring Mission, MERRA-2, NOAA, Freie University Berlin, and HadISST datasets were used. The results of the study showed that significant LSCDs with a 99% confidence level were SSH, SST, LHFOL, SHFOL, T2M, U500, Nino-3.4, and Nino-4. During the test period, compared with MR models of 0.15 to 0.49 and principal components of -0.16 to 0.43, the ANN and BRNN models had better predictive skills with correlation coefficients of 0.40 to 0.74 and 0.34 to 0.70, respectively. It can be concluded that the ANN and BRNN models enable us to predict monthly precipitation in the Balochistan region with lagged LSCDs. Extended Abstract Introduction Pakistan is one of the most vulnerable countries due to climate change. According to the Global Climate Risk Index (GCRI) report, with the passage of time, its vulnerability is increasing. The main threats consist of a rise in temperature, irregular patterns of rainfall, a rise in the sea level, and extreme events such as droughts, floods, and heatwaves. Precipitation plays an important role in the economy of Pakistan as an agricultural country that has experienced variations, particularly in a recent couple of decades when a sharp jump in global atmospheric temperatures was noticed. The southern part of Pakistan which comprises the Sindh and Balochistan provinces has an arid climate and is mostly affected by severe drought events due to less amount of rainfall throughout the year as compared to other parts of the country. This study’s main focus is Balochistan province which has a long history of severe droughts and is an important province in terms of agriculture. Less than 250mm is the annual average rainfall received by the region. It is predicted that throughout the 21st century if adaptation measures are not taken, there will be a continuous increase in droughts and scarcity of water which will adversely affect the lives of people and the economy. Globally, precipitation and droughts are strongly linked with the large-scale climate drivers through atmospheric associations but there are no detailed studies in Balochistan province that consider a wide variety of large-scale climate drivers and their lagged association with precipitation. The main objective of the present study is to identify the significant large-scale climate drivers in the Balochistan province of Pakistan and improve the prediction skill of monthly precipitation by applying the principal component analysis, artificial neural network, Bayesian regularization neural network, and multiple regression analysis considering the lagged association of climate drivers. Methodology In the present study, first, large-scale climate drivers from NOAA HadISST, MERRA-2, and Freie University Berlin in NetCDF files were processed using the ArcGIS model builder along with the precipitation data from GLDAS and TRMM. The data were first normalized in the range of 0 and 1 using the min-max normalization formula. For showing the 12 months’ lagged association between the large-scale climate drivers and precipitation, a cross-correlation method was employed and the heatmaps were created in Origin 2021b software to show significant lagged correlations. Principal component analysis was applied and the variances and Eigenvalues for each of the components were calculated. Multi-layer feed-forward neural network with a back propagation algorithm was used for the prediction of monthly precipitation with the first two PCs in most cases, three and one in some cases as the predictors. Bayesian Regularization Neural Network (BRNN) was applied which is a version of Artificial Neural Networks (ANN) and is a more powerful method as compared to conventional ANN. MATLAB R2015a environment was used for this purpose. Then, Multiple Regression Analysis (MRA) was carried out and was used as a benchmark for comparing ANN and BRNN models using significantly lagged climate indices in one case and in another case the selected PCs. To validate the performance of all developed models, the TRMM dataset was used. Finally, time series graphs and Radar charts were prepared for comparison. Discussions The findings of cross-correlations between monthly precipitation and large-scale climate drivers in Balochistan province showed the most dominant climate drivers as surface specific humidity, sea surface temperature, latent and sensible heat fluxes over land, 500hpa zonal velocity, 2m air temperature, Nino-3.4 and Nino-4, and the highest correlations were noticed for surface specific humidity, sea surface temperature, latent and sensible heat fluxes over land, 500hpa zonal velocity, 2m air temperature. It is noteworthy that in each district, the lagged correlation of maximum climate indices with precipitation was distinct at distinct lags. The multiple regression best models were selected on the basis of no violation of the limits of statistical significance and lower errors. The MLR model’s performance was low in the region where some of the districts had low correlations. The results from ANN and BRNN models showed that the BRNN models were on a lower side than the ANN models with higher values of correlation coefficient showing their capability of finding the pattern and trend of the GLDAS precipitation. The results of the evaluation of the generalization capability of all models on the TRMM dataset showed that both ANN and BRNN models relatively performed well as correlation and error values were closer to each other. Conclusions The results of this study showed that, first, significant LSCDs with 99% confidence level were surface specific humidity (SSH), sea surface temperature (SST), latent and sensible heat fluxes over land (LHFOL and SHFOL), 2m air temperature (T2M), 500hpa zonal velocity (U500), Nino-3.4, and Nino-4. Second, to predict the monthly precipitation using lagged LSCDs and principal components, MR models were developed. MR models’ performance was low. The highest Pearson correlation in MR models during the training set was observed for Musakhel district as 0.65 and during the test as 0.49 for Musakhel and Zhob. In MR-PC models, the highest correlation was recorded for Awaran (0.50) during the training and for Quetta (0.43) during the test. Third, ANN and BRNN models were developed using the selected PC components and gave higher correlations as compared to regression models indicating their capability of finding the pattern and trend of the observations. They generally manifested lower errors and are more reliable for the purpose of prediction in the region. Maximum correlations during training were 0.77 and 0.73 in both ANN and BRNN models, respectively, for the Lasbela district. In the test case, they were 0.74 and 0.70 for Awaran. Their generalization ability was tested on the TRMM dataset. In conclusion, this study divulged the possibility of monthly precipitation prediction using ANN and BRNN and lagged LSCDs for the study region. It is explicit that the response of precipitation to climatic factors is delayed. Comparing the results of the ANN, BRNN, and MLR, it can be concluded that the artificial intelligence approaches such as ANN and BRNN are reliable nonlinear statistical options that can generate similar and in some cases better forecasts and could be useful for the agriculture and water management in the Balochistan region as it is a very important part of Pakistan in terms of agriculture. In the future, this research will be extended by developing the prediction models by including more LSCDs and can be improved by the approaches of genetic algorithm for the best input selection and the use of observed station-based precipitation data sets. [ABSTRACT FROM AUTHOR]

This study deals with the development of artificial intelligence (AI)-based bicycle level of service (BLOS) models for urban road segments carrying heterogeneous traffic. To accomplish this, the required data sets on the road geometric, traffic and built-environmental conditions are collected from 84 road segments located in various parts of four Indian cities. The satisfaction levels of bicyclists at each site are also assessed using a Likert scale of 1–6 (excellent–worst). Subsequently, three promising AI techniques namely, Multivariate adaptive regression splines (MARS), Genetic programming (GP) and Bayesian regularization neural network (BRNN) are utilized to develop the BLOS models. All models are trained and tested with eight significant attributes of the road segments. Among all models, the MARS-based one has shown the best prediction performance in the present context with a coefficient of determination (R2) value of 0.92 with averaged observations. On the other hand, GP has produced the simplest (regression-like) but reliable model, which is the most favourable for field applications. The relative importance of input variables has concluded that the outermost lane width, traffic volume, on-street parking activities and pavement condition index are by far the most important variables in the present context. Hence, these attributes should be largely prioritized in the planning process to enhance the perceived BLOS effortlessly. [ABSTRACT FROM AUTHOR]

Urban intersections in India constitute a significant share of pedestrian fatalities. However, model-based prediction of pedestrian fatalities is still in a nascent stage in India. This study proposes an artificial neural network (ANN) technique to develop a pedestrian fatal crash frequency model at the intersection level. In this study, three activation functions are used along with four different learning algorithms to build different combinations of ANN models. In each of these combinations, the number of neurons in the hidden layer is varied by trial and error method, and the best results are considered. In this way, 12 sets of pedestrian fatal crash predictive models are developed. Out of these, Bayesian Regularization Neural Network consisting of 13 neurons in the hidden layer with 'hyperbolic tangent-sigmoid' activation function is found to be the best-fit model. Finally, based on sensitivity analysis, it is found that the 'approaching speed' of the motorized vehicle has the most significant influence on the fatal pedestrian crashes. 'Logarithm of average daily traffic' (ADT) volume is found to be the second most sensitive variable. Pedestrian-vehicular interaction concerning 'pedestrian-vehicular volume ratio' and lack of 'accessibility of pedestrian cross-walk' are found to be approximately as sensible as 'logarithm of ADT'. [ABSTRACT FROM AUTHOR]

The application of remote sensory images in crop monitoring has been increasing in the recent years due to its high classification accuracy. In this paper, a novel parallel classification methodology is proposed using a new clustering and classification concept. A novel neural network model with the Bs-Lion training algorithm is developed by integrating the Bayesian regularization training with the Lion Algorithm. Here, two levels of parallel processing are performed, namely parallel WLI-Fuzzy clustering and parallel BS-Lion neural network classification. The experimentation of the proposed parallel methodology is carried out using satellite images obtained from the Indian remote sensing satellite IRS-P6. The performance of the proposed system is compared with the existing techniques using validation measures accuracy, sensitivity and specificity. The experimentations resulted in promising results with an accuracy of 0.8994, sensitivity of 0.8682 and specificity of 0.8739, which favour the performance of the proposed parallel architecture in the classification. [ABSTRACT FROM AUTHOR]

Quantitative structure-activity relationship methods are used to study the quantitative structure triboability relationship (QSTR), which refers to the tribology capability of a compound from the calculation of structure descriptors. Here, we used the Bayesian regularization neural network (BRNN) to establish a QSTR prediction model. Two-dimensional (2D) BRNN-QSTR models can flexibly and easily estimate lubricant-additive antiwear properties. Our results show that electron transfer and heteroatoms (such as S, P, O, and N) in a lubricant-additive molecule improve the antiwear ability. We also found that molecular connectivity indices are good descriptors of 2D BRNN-QSTR models. [ABSTRACT FROM AUTHOR]

With growing urbanization, it has become necessary to manage this growth smartly. Specifically, increased electrical energy consumption has become a rapid urbanization trend in China. A building model based on a neural network was proposed to overcome the difficulties of analytical modelling. However, increased amounts of data, repetitive computation, and training time become a limitation of this approach. A simplified model can be used instead of the full order model if the performance is acceptable. In order to select effective data, Mean Impact Value (MIV) has been applied to select meaningful data. To verify this neural network method, we used real electricity consumption data of a shopping mall in China as a case study. In this paper, a Bayesian Regularization Neural Network (BRNN) is utilized to avoid overfitting due to the small amount of data. With the simplified data set, the building model showed reasonable performance. The mean of Root Mean Square Error achieved is around 10% with respect to the actual consumption and the standard deviation is low, which reflects the model's reliability. We also compare the results with our previous approach using the Levenberg–Marquardt back propagation (LM-BP) method. The main difference is the output reliability of the two methods. LM-BP shows higher error than BRNN due to overfitting. BRNN shows reliable prediction results when the simplified neural network model is applied. [ABSTRACT FROM AUTHOR]

The compression index (Cc) is a necessary parameter for the settlement calculation of clays. However, determination of the compression index from oedometer tests takes a relatively long time and leads to a very demanding experimental working program in the laboratory. Therefore, geotechnical engineering literature involves many studies based on indirect methods such as multiple regression analysis (MLR) and soft computing methods to determine the compression index. This study is aimed to predict the compression index by using extreme learning machine (ELM), Bayesian regularization neural network (BRNN), and support vector machine (SVM) methods. The selected variables for each method are the natural water content (wn), initial void ratio (e0), liquid limit (LL), and plasticity index (PI) of clay samples. Many trials were carried out in order to get the best prediction performance with each model. The application results obtained from the models were also compared based on the correlation coefficient (R), coefficient of efficiency (E), and mean squared error (MSE). The results indicate that the BRNN method has better success on estimation of the compression index compared to the ELM and SVM methods. [ABSTRACT FROM AUTHOR]