جستجوی مقالات مرتبط با کلیدواژه « support vector machine » در نشریات گروه « عمران »

Traffic crashes are a significant problem in low and middle-income countries, while there is a worrying trend of increasing fatal and injury crashes Iran. This highlights the urgent need to analyze the causes of such accidents to improve road safety and reduce their negative consequences. To address this issue, a study was conducted to investigate the factors that contribute to the severity of rural crashes in Zanjan province, using advanced machine learning models such as Support Vector Machine and Decision Tree. The study utilized a crash database of 25,000 incidents over a 9-year period, and after cleaning the data, the models were developed in Python. The findings suggest that “type of crash”, “at-fault driver's vehicle type”, and “kilometer occurrence of the crash” are key variables for predicting the severity of these crashes. The Decision Tree model was also found to be more accurate than the Support Vector Machine model, particularly in predicting severe crashes. This study provides valuable insights for improving road safety and reducing the harmful effects of traffic crashes in rural areas.

Every country relies on soil as a vital natural resource that significantly contributes to environmental conservation and food production. Preparation of soil nutrient distribution map serves as a valuable tool for managers to make decisions. Due to the time-consuming and expensive nature of laboratory analysis for these variables on a large scale, efforts have been made to explore soil nitrogen through remote sensing. The current research deals with the application of remote sensing methods along with regression and random forest models to predict total soil nitrogen in Gilan province. This study aimed to answer two main questions: (1) Can SAR data be used to quantify total soil nitrogen (TSN) (2) How do SVM, BRT and RF algorithms perform in predicting soil nitrogen content?

This study focused on evaluating the data capabilities of Landsat-9 and Sentinel-1 satellites individually and in combination, using advanced algorithms such as Support Vector Machine (SVM), Boosted Regression Tree (BRT), and Random Forest (RF). The purpose of this evaluation was strategic, aiming to showcase the diverse conditions of the study area based on land cover/land use, climatic, and topographical parameters. Various variables, including climate parameters, topographic components, and remote sensing subscale indices, were investigated in conjunction with SAR data and optical images. Nonlinear machine learning algorithms, specifically SVM, RF, and BRT, were employed to predict total soil nitrogen status by modeling complex relationships between soil properties and environmental variables. R software, utilizing the CARET package for parameter input, was employed to implement the algorithm.

The results indicated the following: RF and BRT algorithms outperformed SVM and were effective in monitoring total soil nitrogen values. Multi-temporal SAR images showed higher accuracy in monitoring total soil nitrogen content compared to optical remote sensing data, facilitating more realistic predictions in paddy soils. The integration of environmental variables led to an increase in the accuracy of algorithms, where remote sensing variables played a crucial role, contributing to 61% and 51% effects in RF and BRT algorithms, respectively. The comparison of SVM and RF algorithms revealed that RF ranked second after the BRT algorithm, and the accuracy of total soil nitrogen estimation was not achieved with the SVM algorithm. However, both BRT and RF algorithms were able to monitor changes in total soil nitrogen. BRT performed better, accurately recording 58% of changes, as evidenced by a higher R2 value (0.58) and lower RMSE (0.25 mg/kg) and MAE (0.19 mg/kg) values.

In conclusion, the following key points were extracted from this research: 1) RF and BRT algorithms outperformed SVM in effectively monitoring total soil nitrogen levels; 2) multi-temporal SAR images demonstrated higher accuracy in tracking total soil nitrogen compared to optical remote sensing, enabling precise predictions in paddy soils; 3) the incorporation of environmental variables enhanced algorithmic accuracy; and 4) remote sensing variables contributed 61% and 51% to RF and BRT algorithms, respectively.

The downstream scour of the vertical drop can be one of the causes of instability and failure of this structure. In the present study, the downstream scour depth of this structure predicted using the support vector machine (SVM) method. For this purpose, 104 experimental data used to estimate the scour depth. hese data are a function of the two dimensionless parameters of dansimetric Froude number (Frj) and tailwater depth (yt / yj) that have been entered into the SVM in three different models. To evaluate the results, the evaluation criteria of R2, NRMSE, DC, and MARE used. The results showed that model number (1) with the input combination (Frj and yt / yj) with R2 = 0.9777, DC = 0.929, NRMSE = 0.0775, and MARE = 11.89% for the test stage leads to the best result. The SVM method also has appropriate accuracy, acceptable results, and desirable performance in estimating the scour depth. Also, it was found that the densimetric froude number has a greater effect on estimating the relative scour depth compared to the tailwater depth.

One of the information needed for all planning problems and specifically transportation planning is to have accurate prediction about the future. Traffic variables prediction is one of the efficient tools in travel demand management. Using this tool and advanced traveler information systems (ATIS), the predicted traffic variables are informed to the users and transportation system operators to make plans and set policies. In this study, the average speed and traffic volume of the Karaj to Chalus suburban road with the high variation of traffic variables in the north of Iran is predicted. The Karaj to Chalous road is part of the route from Tehran as the capital of Iran to the country's northern coast. Along the Karaj to Chalous road, three parallel roads, with different lengths, connect Tehran with the cities of the north. In general, finding the pattern of non-mandatory trips is more complicated than mandatory trips. Generally, the predictive methods are divided into three groups, naïve, parametric and non-parametric methods. Among the various predictive models, the SARIMA as a parametric model and the artificial neural network and the support vector machine as nonparametric models are employed. In the data pre-processing step, the variables affecting the average speed and traffic volume are extracted and added to the dataset as predictor variables. These variables are related to time, calendar, holidays, weather, and roads blockage. Also, because of the importance of the maximum and minimum values of traffic speed and volume, as critical values and rare events, models are evaluated with emphasis on the prediction of rare events compared to normal values. The results show that, for the test data, the lowest root mean square error of predicting the average traffic speed and traffic volume are obtained using artificial neural network and support vector machine models equals 139 vehicles per hour and 5 kilometers per hour, respectively. In terms of R2 of prediction-observation plot, the performance of SARIMA for predicting the average speed and traffic volume is the same for the test dataset. In contrast the R2 of hourly traffic volume prediction is higher for the training data. The R2 of artificial neural network model and the support vector machine for traffic volume prediction is higher than traffic speed prediction. The lowest root mean square error of predicting the first and fourth quartile of the observed average traffic speed values is obtained by support vector machine models and artificial neural network, respectively. Also, predicting the first quartile and fourth quartile of the observed traffic volume values by the support vector machine model is more accurate than two other models. Using predicted traffic parameters and providing them to travelers and transportation agencies by intelligent transportation systems leads to make a balance between travel demand and travel supply in the near future which is the main aim of this study. Travelers can have a better personal plan for their future trips based on these predictions. Also, the transportation agencies are more prepared to deal with critical traffic situations and can prevent traffic congestion.

In predicting the behavior of geotechnical structures that are constructed on unsaturated soils, knowing the geomechanical properties, especially effective stress, is of particular importance. The effect of suction in unsaturated soils changes the volumetric and shear behavior of soil. Therefore, in determining the shear strength in unsaturated soils, the effective stress parameter (χ) plays an essential role as a function of suction. Determining this parameter requires a lot of time and money that is spent on laboratory tests. The purpose of this study is to evaluate several intelligent methods for creating models that indirectly accurately estimate parameter (χ). For this purpose, 120 data (which are the results of thriaxial, shear, pressure plate and filter paper tests) and intelligent methods of random forest, support vector machine and k-nearest neighbor were used in WEKA intelligent software. The results show that the models developed by the three intelligent methods of random forest, support vector machine and k-nearest neighbor, have excellent performance and accuracy. But the random forest model is better than the other two models in evaluating the test data with R2 = 0.918 and RMSE = 0.079. In addition, in this study, sensitivity analysis was performed to determine the importance of the parameters affecting parameter (χ). Among the input parameters in modeling, it was found that the volume water content parameter (θ_r/θ_s ) has the greatest effect on parameter (χ).

Predicting traffic variables and informing the passengers and the transportation network operators is one way to manage the travel demand. By informing the future state of traffic through intelligent transportation systems, there is more readiness to avoid congestion. In this study, three machine learning models, including support vector machine (SVM), random forest (RF), and long short-term memory (LSTM), were used to predict the hourly traffic state, consist of light, semi-heavy and heavy states, for Karaj to Chalus rural road in the north of Iran. Predictor variables of mid-term models are calendar information, weather, and road blockage policies. In contrast, in short-term models, in addition to mentioned variables, the observed traffic states in the past three to eight hours have been used, and these models can only predict the future of one and two hours of future. The results show that short-term LSTM is the most accurate traffic state predictor model with accuracy equal to 90.11%. Among the mid-term models, the LSTM model has predicted traffic state more accurately than SVM and Rf, and its accuracy is equal to 82.07%. Also, LSTM has the highest values of f1 to predict light, semi-heavy, and heavy, which are equivalent to 0.86, 0.93, and 0.81, respectively. Also, the hour, holiday, and type of holiday variables and traffic state observed in 3 to 8 hours later variables have the greatest effect on increasing the accuracy of mid-term and short-term models, respectively.

Weirs are structures to provide the passage for excess water to flow from upstream to downstream in flood conditions. Since the special geometry of ogee weir in terms of matching the flow Trajectory with the surface of the weir, increases the efficiency of this type of weirs, so the use of terminal structures to dissipate destructive energy downstream of weir is of great importance. Increasing the water level in the Ogee weirs generally increases the contact level and head loss, but specifically in gabion weirs leads to increased permeability. One of the new methods of energy dissipation is the use of gabion structures, artificial roughness, blocks and lattice plates in the flow path as alternative solutions to using the stilling basins. Based on the results of previous research in the field of artificial intelligence, in the present study, the amount of energy consumption of the Ogee weir terminal structure (gabion structure) was predicted using the support vector machine and the effect of dimensions and grain size of the gabion structure. The amount of energy dissipation was also examined.

For this purpose, the experiments were performed in a rectangular channel located in the hydraulic laboratory of Maragheh University, 13 meters long, 120 cm wide and 80 cm high, with a metal floor and a glass wall 1 cm thick, which allows accurate observation of flow behaviors. The gabion used in this experiment was made of rebar number 6 with a width of 120 cm, a length of 10 cm (in the direction of flow) and a variable height, then it was surrounded by a metal mesh with a thickness of 1 mm and a diameter of 1 cm. To predict energy dissipation in the support vector machine, we need a series of functions based on the parameters extracted in the dimensional analysis. Models with different percentages of training and testing (65-35, 70-30, 75-25 and 80-80) and using the radial basis function (RBF), with the appropriate gamma value obtained during trial and error, were checked by a support vector machine. The following criteria were used to evaluate the obtained results and evaluate the efficiency of the models. 1- Normal root mean square error (RMSE), which no matter how close the index (RMSE) is to zero, the model has high accuracy. 2. The normal root mean square error (NRMSE) where NRMSE below 10% indicates the accuracy of the model, 10-20% indicates the suitability of the model, 20-30% the average accuracy and more than 30% indicates the weakness of the model. 3. Performance coefficient (Nash and Sutcliffe) which shows the linear correlation between the measured and predicted values and the closer the value is to one, the better the data correlation.

The results are presented in three sections: Laboratory, Soft Computing and Sensitivity Analysis. To investigate the energy dissipation of the downstream gabion structure, the Ogee weir was performed according to the various variables extracted in the dimensional analysis, including the number and width of openings, grain size, and Froude number. It can be seen that in all models, energy dissipation values are directly related to the Froude number. At the same flow rate, with decreasing Froude number, the flow depth increases and most of the flow passes through the gabion structure as a weir and the energy dissipation decreases. At shallow flow where all or most of the flow passes through the structure, most of the energy dissipation is due to the collision of the flow with the particles inside the structure. However, in higher Froude numbers, the structure is submerged under the flow and both internal flow and overflow are effective in energy dissipation, with the difference that in overgrowth structures, overflow and in fine-grained structures, internal flow is predominant.

Therefore, at a constant opening, fine-grained has the highest amount of energy dissipation. In other words, with increasing the diameter of the rock grains, the volume of the pores increases and the flow passes more easily through the rock grains and the turbulence and flow of the flow in this area decreases, so the energy dissipation decreases. It can be seen that for all models, the values predicted by the support vector machine are close to the laboratory results, but in the Froude numbers the accuracy of the support vector machine is further reduced. The reason for this can be considered as turbulence and turbulence in higher Froude numbers. Finally, it can be stated that the energy dissipation of the gabion structure in the test and training phase has acceptable compliance and overlap with laboratory values. Sensitivity analysis showed that the parameter of relative water depth after gabion has the greatest effect on the correct prediction of energy consumption due to gabion structure.

In recent decades, the science of structural health monitoring has played a key role in preventing damage and extending the life of structures. To conduct behavioral assessment, it is desirable to use tools that achieve sufficient accuracy with low cost. The processing of behavioral data requires methods that are able to identify and correctly troubleshoot different levels of damage from existing information. Nowadays, sensors are used to measure the behavior of structures including deformations and displacements and even deflections, but these sensors have some weak points. For example, Risk of damage to the sensor, pointwise and one-dimensional measuring, their data is difficult to analyze and using multiple or high-tech sensors becomes expensive. Optical behavior measurement and close-range photogrammetric operations have recently received attention due to their low cost and good accuracy. This method has some advantages like Indirect contact with objects, high-speed image capture, easy access to convenient digital cameras, low viewing costs, and the ability to process composite and instant data with easy operation. In addition, the high flexibility of this method in measuring accuracy and design capability to achieve predetermined accuracy is an important feature of this tool. Analytical methods are based on rules or equations that provide a clear definition of the problem. These methods work well in the cases which the rules are accurately clear and defined but there are many practical cases for which the rules are not known or it is very difficult to discover that calculations cannot be performed using analytical methods. Neural network is a generalizable model, which is based on the experience of a set of training data and therefore free of explicit law. Neural networks have the ability to collect, store, analyze, and process large amounts of data from numerical analyzes or experiments. Therefore, they have the ability to predict and build diagnostic models to solve various engineering problems and tasks In this paper, an attempt has been made to use this method to measure and troubleshoot laboratory model of a scaled suspension bridge that has a relatively complex behavior. For this purpose, the structure was subjected to uniform static loading in three step levels with three states: healthy and damaged in the deck and cables. Damages were created quite intentionally in the laboratory model, and from the information obtained, a database of bridge behavior in various situations was created. In order to assess the feasibility of using different methods in data processing and troubleshooting, first the data in the database were used in a simple linear method (direct comparison) and training in algorithms of machine learning methods. After that, deliberate damage was done again in the laboratory structure to allow testing the efficiency and accuracy of different methods. Finally, the accuracy, precision, and stability of the data processing methods of the support vector machine and artificial neural network were compared. The results showed that by object bundle justification of two-dimensional optical behaver measurement with close-range photogrammetry, a guaranteed accuracy of 0.0021 mm could be achieved. Using intensity image processing seems helpful to ease the calculation. Using high number of nodes in hidden layer makes it more difficult and time-consuming to train the neural network. In the first level of processing, the detection of the presence or absence of damage was associated with the complete superiority of neural networks with 100% accuracy and in the second level, the detection of the affected area, depending on the type of processing, the neural network with hyperbolic tangent transfer function archived 93% accuracy and the support vector machine archived 68% of the accuracy.

Rockburst in deep underground excavations is a phenomenon that is observed in the form of sudden rock failure and release of strain energy stored in underground mines and rock tunnels, usually in stressful places and at a depth greater than the earth's surface. Due to this explosive failure, rock in pieces Small and large are scattered around and cause damage to humans or equipment. In the present study, the potential of rockburst using data mining techniques and comparing their results with three experimental methods, tangential stress criterion (SC), brittleness criterion (BC) and elastic energy index (EEI) are investigated. For this purpose, data mining models such as support vector machine (SVM), k-nearest neighbor (KNN), Bayesian networks (BNs), artificial neural network (ANN) and CHAID tree model have been used in WEKA software. The results show the superiority of data mining algorithms over experimental methods. Among data mining algorithms, the support vector machine model had the highest accuracy with 80.8% accuracy. However, measuring the accuracy of the models by RMSE method considers the artificial neural network model with the value of RMSE = 0.337 as the best model. In addition, among the experimental methods, the elastic energy index method with an accuracy of 60.7 is the best method.

Today, structural health monitoring techniques that can detect damage in early stages have become very important. For these reason, such techniques must be able to detect slight damages. However, a large number of algorithms that have proposed so far for damage detection are unable to identify early-stage damages. One of the approaches that can be employed is multifractal method. In this paper, a damage detection technique is proposed based on multifractal detrend fluctuation analysis and blind source separation. In the first stage, the accuracy of three methods of blind source separation is compared and the most efficient method in decomposing structural vibration signals is selected. These methods include blind modal identification, combined method, and sparse coding. Three structural models are employed to investigate efficiency of the procedures which consists of a range of numerical MDOF model with limited degree of freedom to real structures. In the second stage, a damage index is proposed based on the width of multifractal spectrum. Results show that the aforementioned method is able to identify various damage patterns and can detect slight damages.

Due to the excessive concern about environmental issues, researchers had to come up with a better solution to control the Wastewater treatment plants (WWTPs).In this research, two approaches, including Artificial Neural Network (ANN) and Support Vector Machine (SVM) have been used for modeling the effluent quality of the Tabriz Wastewater Treatment Plant. Input data of models consist ofBODinf, CODinf, TSSinf, and PHinf of influent sewage related to Tabriz Treatment Plant which has been used to predict the corresponding value of BODeff, CODeff, and TSSeff concerning the treatment plant effluent. The daily, weekly, and monthly average data have been studied. According to the results, the two approaches mentioned, have the best performance in the prediction of the monthly average dataset of effluent parameters of Tabriz Wastewater Treatment Plant.

Knowing the tree species combination of forests provides valuable information for studying the forest’s economic value, fire risk assessment, biodiversity monitoring, and wildlife habitat improvement. Fieldwork is often time-consuming and labor-required, free satellite data are available in coarse resolution and the use of manned aircraft is relatively costly. Recently, unmanned aerial vehicles (UAV) have been attended to be an easy-to-use, cost-effective tool for the classification of trees. In fact, given the cost-efficient nature of UAV derived SfM, coupled with its ease of application, it became a popular choice. The type of imagery is an important factor in classification analysis because the spatial and spectral resolution can influence the accuracy of classification. On the other hand, classification algorithms also play an important role in the accuracy of tree species identification. So, this study investigated the performance of four classifiers for tree species classification using UAV-based high-resolution imagery in broadleaf forests and takes a comparative approach to examine the three non-parametric classifiers including support vector machines (SVM), random forest (RF), artificial neural network (ANN), and one parametric classifier including linear discriminant analysis (LDA) classifiers in heterogeneous forests of Noor city located in Mazandaran province. In June 2019, the study area was photographed. The field survey was carried out to record the species and position of the mature overstory trees which were clearly identifiable on the orthomosaics. Individual tree crowns were clipped by one-meter buffer and the digit numbers were summarized at for each tree by computing descriptive statistics from the orthomosaics. Using zonal statistics, mean, standard deviation, variance, unique, range, mode, and median were calculated for raw bands (Red, Green, Blue), vegetation indices (NRB, NGB), and band ratios (G/R, R/B) from RGB orthomosaics. We classified the tree into 4 classes: Parrotia persica (Ironwood tree), Populus capsica (Caspian poplar), Ulmus minor (Common Elm), and Quercus castaneifolia (Chestnut-leaved oak). Finally, the classification algorithms were applied using R software. The classification accuracy for identified trees was performed using 10-fold cross-validation by computing the producer’s accuracy, user’s accuracy, and Overall accuracy. All algorithms resulted in overall accuracies above 80%. Of course, the results showed that, as a parametric algorithm, LDA with an overall accuracy of 0.87 provided the best results for tree classification, because it does not require the tuning of free parameters. As for parameter value, the mean was the most important that this can be related to the similarity of this feature in any sample. Caspian poplar with user accuracy of 0.97 and Ironwood tree with user accuracy of 0.72 had the highest and lowest classification accuracy, respectively. Caspian poplar high accuracy is probably due to its crown color which is quite different from the other species. The main error (misclassification) is a classification between “Ironwood tree” and “Common Elm” classes. This may be caused by the fact that the spectral signatures between Ironwood tree and Common Elm trees are very similar. In general, our study showed that UAV derived orthomosaic can be used for tree classification with very high accuracy in mix broadleaf forests by different algorithms.

Today, it is possible to detect damage in the early stages with the aid of structural health monitoring (SHM) techniques to avoid financial losses and loss of lives. However, large expenses of SHM systems has caused low popularity of such systems in our country. The aim of this study is to provide a low-cost damage detection technique for bridges based on signal processing and machine learning. To reduce expenses, the number of sensors to monitor vibration of the structure is decreased to only one sensor. Since reduction of number of sensors can lead to drop in damage detection accuracy, most up to date signal processing methods are used. In the first step of the paper, several time-frequency signal processing techniques are compared and EWT is selected as the best signal processing method. In the next step, after decomposition of signals by time-frequency techniques, a new damage index is introduced base on cross wavelet transform (CWT) and then calculated damaged indices are classified using support vector machine (SVM) to be able to distinguish healthy and damage states. Results show that the proposed method can detect damage with high accuracy.