Groundwater level prediction of landslide based on classification and regression tree

According to groundwater level monitoring data of Shuping landslide in the Three Gorges Reservoir area, based on the response relationship between influential factors such as rainfall and reservoir level and the change of groundwater level, the influential factors of groundwater level were selected. Then the classification and regression tree（CART） model was constructed by the subset and used to predict the groundwater level. Through the verification, the predictive results of the test sample were consistent with the actually measured values, and the mean absolute error and relative error is 0.28 m and 1.15%respectively. To compare the support vector machine（SVM） model constructed using the same set of factors, the mean absolute error and relative error of predicted results is 1.53 m and 6.11% respectively. It is indicated that CART model has not only better fitting and generalization ability, but also strong advantages in the analysis of landslide groundwater dynamic characteristics and the screening of important variables. It is an effective method for prediction of ground water level in landslides.

Prediction method of restoring force based on online AdaBoost regression tree algorithm in hybrid test

In order to solve the poor generalization ability of the back-propagation(BP)neural network in the model updating hybrid test,a novel method called the AdaBoost regression tree algorithm is introduced into the model updating procedure in hybrid tests.During the learning phase,the regression tree is selected as a weak regression model to be trained,and then multiple trained weak regression models are integrated into a strong regression model.Finally,the training results are generated through voting by all the selected regression models.A 2-DOF nonlinear structure was numerically simulated by utilizing the online AdaBoost regression tree algorithm and the BP neural network algorithm as a contrast.The results show that the prediction accuracy of the online AdaBoost regression algorithm is 48.3%higher than that of the BP neural network algorithm,which verifies that the online AdaBoost regression tree algorithm has better generalization ability compared to the BP neural network algorithm.Furthermore,it can effectively eliminate the influence of weight initialization and improve the prediction accuracy of the restoring force in hybrid tests.

A New Approach to Predict Financial Failure： Classification and Regression Trees （CART）

The increase of competition, economic recession and financial crises has increased business failure and depending on this the researchers have attempted to develop new approaches which can yield more correct and more reliable results. The classification and regression tree （CART） is one of the new modeling techniques which is developed for this purpose. In this study, the classification and regression trees method is explained and tested the power of the financial failure prediction. CART is applied for the data of industry companies which is trade in Istanbul Stock Exchange （ISE） between 1997-2007. As a result of this study, it has been observed that, CART has a high predicting power of financial failure one, two and three years prior to failure, and profitability ratios being the most important ratios in the prediction of failure.

Mapping species assemblages of tropical forests at different hierarchical levels based on multivariate regression trees

Background: Vegetation distribution maps are of great significance for nature protection and management. In diverse tropical forests, accurate spatial mapping of vegetation types is challenging;the high species diversity and abundance of rare species challenge classification concepts, while remote sensing signals may not vary systematically with species composition, complicating the technical capability for delineating vegetation types in the landscape.Methods: We used a combination of field-based compositional data and their relations to environmental variables to predict the distribution of forest types in the Wuzhishan National Natural Reserve(WNNR), Hainan Island,China, using multivariate regression trees(MRT). The MRT was based on arboreal vegetation composition in 132plots of 20 m×20 m with a regular spacing of 1 km. Apart from the MRT, non-metric multidimensional scaling(NMDS) was used to evaluate vegetation-environment relationships.Results: The MRT model worked best when using 14 key environmental variables including topography, climate,latitude and soil, although the difference with the simpler model including only topographical variables was small. The full model classified the 132 plots into 3 vegetation types, 6 formation groups, 20 formations and 65associations at different hierarchical syntaxonomic levels. This model was the basis for forest vegetation maps for the WNNR. MRT and NMDS showed that elevation was the main driving force for the distribution of vegetation types and formation groups. Climate, latitude, and soil(especially available P), together with topographic variables, all influenced the distribution of formations and associations.Conclusions: While elevation determines forest-type distributions, lower-level syntaxonomic forest classes respond to the topographic diversity typical for mountains. Apart from providing the first detailed forest vegetation map for any part of WNNR, we show how, in spite of limitations, MRT with existing environmental data can be a useful method for mapping diverse and remote tropical forests.

Predicting plant disease epidemics using boosted regression trees

Plant epidemics are often associated with weather-related variables.It is difficult to identify weather-related predictors for models predicting plant epidemics.In the article by Shah et al.,to predict Fusarium head blight(FHB)epidemics of wheat,they explored a functional approach using scalar-on-function regression to model a binary outcome(FHB epidemic or non-epidemic)with respect to weather time series spanning 140 days relative to anthesis.The scalar-on-function models fit the data better than previously described logistic regression models.In this work,given the same dataset and models,we attempt to reproduce the article by Shah et al.using a different approach,boosted regression trees.After fitting,the classification accuracy and model statistics are surprisingly good.

Artificial Neural Network （ANN） and Regression Tree （CART） applications for the indirect estimation of unsaturated soil shear strength parameters

The shear strength parameters of soil （cohesion and angle of internal friction） are quite essential in solving many civil engineering problems. In order to determine these parameters, laboratory tests are used. The main objective of this work is to evaluate the potential of Artificial Neural Network （ANN） and Regression Tree （CART） techniques for the indirect estimation of these parameters. Four different models, considering different combinations of 6 inputs, such as gravel %, sand %, silt %, clay %, dry density, and plasticity index, were investigated to evaluate the degree of their effects on the prediction of shear parameters. A performance evaluation was carried out using Correlation Coefficient and Root Mean Squared Error measures. It was observed that for the prediction of friction angle, the performance of both the techniques is about the same. However, for the prediction of cohesion, the ANN technique performs better than the CART technique. It was further observed that the model considering all of the 6 input soil parameters is the most appropriate model for the prediction of shear parameters. Also, connection weight and bias analyses of the best neural network （i.e., 6/2/2） were attempted using Connec- tion Weight, Garson, and proposed Weight-bias approaches to characterize the influence of input variables on shear strength parameters. It was observed that the Connection Weight Approach provides the best overall methodology for accurately quantifying variable importance, and should be favored over the other approaches examined in this study.

Simulation of reservoir outflows using regression tree and support vector machine

Water stored in reservoirs has a lot of crucial function,including generating hydropower,supporting water supply,and relieving lasting droughts.During floods,water deliveries from reservoirs must be acceptable,so as to guarantee that the gross volume of water is at a safe level and any release from reservoirs will not trigger flooding downstream.This study aims to develop a well-versed assessment method for managing reservoirs and pre-releasing water outflows by using the machine learning technology.As a new and exciting AI area,this technology is regarded as the most valuable,time-saving,supervised and cost-effective approach.In this study,two data-driven forecasting models,i.e.,Regression Tree(RT)and Support Vector Machine(SVM),were employed for approximately 30 years’hydrological records,so as to simulate reservoir outflows.The SVM and RT models were applied to the data,accurately predicting the fluctuations in the water outflows of a Bhakra reservoir.Different input combinations were used to determine the most effective release.For cross-validation,the number of folds varied.It is found that quadratic SVM for 10 folds with seven different parameters would give the minimum RMSE,maximum R2,and minimum MAE;therefore,it can be considered as the best model for the dataset used in this study.

A retinal blood vessel extraction algorithm based on CART decision tree and improved AdaBoost

This paper presents a supervised learning algorithm for retinal vascular segmentation based on classification and regression tree (CART) algorithm and improved adptive bosting (AdaBoost). Local binary patterns (LBP) texture features and local features are extracted by extracting,reversing,dilating and enhancing the green components of retinal images to construct a 17-dimensional feature vector. A dataset is constructed by using the feature vector and the data manually marked by the experts. The feature is used to generate CART binary tree for nodes,where CART binary tree is as the AdaBoost weak classifier,and AdaBoost is improved by adding some re-judgment functions to form a strong classifier. The proposed algorithm is simulated on the digital retinal images for vessel extraction (DRIVE). The experimental results show that the proposed algorithm has higher segmentation accuracy for blood vessels,and the result basically contains complete blood vessel details. Moreover,the segmented blood vessel tree has good connectivity,which basically reflects the distribution trend of blood vessels. Compared with the traditional AdaBoost classification algorithm and the support vector machine (SVM) based classification algorithm,the proposed algorithm has higher average accuracy and reliability index,which is similar to the segmentation results of the state-of-the-art segmentation algorithm.

Machine learning prediction of methane,ethane,and propane solubility in pure water and electrolyte solutions:Implications for stray gas migration modeling

Hydraulic fracturing is an effective technology for hydrocarbon extraction from unconventional shale and tight gas reservoirs.A potential risk of hydraulic fracturing is the upward migration of stray gas from the deep subsurface to shallow aquifers.The stray gas can dissolve in groundwater leading to chemical and biological reactions,which could negatively affect groundwater quality and contribute to atmospheric emissions.The knowledge oflight hydrocarbon solubility in the aqueous environment is essential for the numerical modelling offlow and transport in the subsurface.Herein,we compiled a database containing 2129experimental data of methane,ethane,and propane solubility in pure water and various electrolyte solutions over wide ranges of operating temperature and pressure.Two machine learning algorithms,namely regression tree(RT)and boosted regression tree(BRT)tuned with a Bayesian optimization algorithm(BO)were employed to determine the solubility of gases.The predictions were compared with the experimental data as well as four well-established thermodynamic models.Our analysis shows that the BRT-BO is sufficiently accurate,and the predicted values agree well with those obtained from the thermodynamic models.The coefficient of determination(R2)between experimental and predicted values is 0.99 and the mean squared error(MSE)is 9.97×10^(-8).The leverage statistical approach further confirmed the validity of the model developed.

Building a Tree Adjusted Logistic Classification Model in Biomarker Data Analyses

Researchers in bioinformatics, biostatistics and other related fields seek biomarkers for many purposes, including risk assessment, disease diagnosis and prognosis, which can be formulated as a patient classification. In this paper, a new method of using a tree regression to improve logistic classification model is introduced in biomarker data analysis. The numerical results show that the linear logistic model can be significantly improved by a tree regression on the residuals. Although the classification problem of binary responses is discussed in this research, the idea is easy to extend to the classification of multinomial responses.

Data-driven methods for predicting the representative temperature of bridge cable based on limited measured data

Cable-stayed bridges have been widely used in high-speed railway infrastructure.The accurate determination of cable’s representative temperatures is vital during the intricate processes of design,construction,and maintenance of cable-stayed bridges.However,the representative temperatures of stayed cables are not specified in the existing design codes.To address this issue,this study investigates the distribution of the cable temperature and determinates its representative temperature.First,an experimental investigation,spanning over a period of one year,was carried out near the bridge site to obtain the temperature data.According to the statistical analysis of the measured data,it reveals that the temperature distribution is generally uniform along the cable cross-section without significant temperature gradient.Then,based on the limited data,the Monte Carlo,the gradient boosted regression trees(GBRT),and univariate linear regression(ULR)methods are employed to predict the cable’s representative temperature throughout the service life.These methods effectively overcome the limitations of insufficient monitoring data and accurately predict the representative temperature of the cables.However,each method has its own advantages and limitations in terms of applicability and accuracy.A comprehensive evaluation of the performance of these methods is conducted,and practical recommendations are provided for their application.The proposed methods and representative temperatures provide a good basis for the operation and maintenance of in-service long-span cable-stayed bridges.

Sustainable Intensification and Large-scale Operation of Cultivated Land Use at the Farmers’ Scale:A Case Study of Shandong Province,China

Sustainable intensification of cultivated land use(SICLU) and large-scale operations(LSO) are widely acknowledged strategies for enhancing agricultural performance.However,the existing literature has faced challenges in precisely defining SICLU and constructing comprehensive indicators,which has hindered the exploration of factors influencing LSO within the SICLU framework.To address this gap,we integrated self-efficacy theory into the design of an index framework for evaluating SICLU.We subsequently employed econometric models to analyze the significant factors that impact LSO.Our findings reveal that SICLU can be divided into four key dimensions:intensive management,efficient output,resource conservation,and ecological environment optimization.Furthermore,it is crucial to incorporate belief-based cognitive factors into the index system,as farmers’ understanding of fertilizer and pesticide application significantly influences their willingness to engage in LSO.Moreover,we identify grain market turnover as the most influential factor in promoting LSO,with single-factor contribution rates reaching 70.9% for cultivated land transfer willingness and 62.5% for the total planting areas.Interestingly,unlike irrigation and agricultural machinery inputs,increased labor inputs correspond to larger planting areas for farmers.This trend may be attributed to reduced labor availability because of rural labor migration,whereas the reduction in irrigation and agricultural input is contingent on innovations in production practices and the transfer of cultivated land management rights.Importantly,SICLU dynamically influences LSO,with each index related to SICLU having an optimal range that fosters LSO.These insights offer valuable guidance for policymakers,emphasizing farmers as their central focus,with the adjustment of input and output factors as a means to achieve LSO as the ultimate goal.In conclusion,we propose research avenues for further enriching the SICLU framework to ensure that it aligns with the specific characteristics of regional agricultural development.

Machine Learning-Driven Classification for Enhanced Rule Proposal Framework

In enterprise operations,maintaining manual rules for enterprise processes can be expensive,time-consuming,and dependent on specialized domain knowledge in that enterprise domain.Recently,rule-generation has been automated in enterprises,particularly through Machine Learning,to streamline routine tasks.Typically,these machine models are black boxes where the reasons for the decisions are not always transparent,and the end users need to verify the model proposals as a part of the user acceptance testing to trust it.In such scenarios,rules excel over Machine Learning models as the end-users can verify the rules and have more trust.In many scenarios,the truth label changes frequently thus,it becomes difficult for the Machine Learning model to learn till a considerable amount of data has been accumulated,but with rules,the truth can be adapted.This paper presents a novel framework for generating human-understandable rules using the Classification and Regression Tree(CART)decision tree method,which ensures both optimization and user trust in automated decision-making processes.The framework generates comprehensible rules in the form of if condition and then predicts class even in domains where noise is present.The proposed system transforms enterprise operations by automating the production of human-readable rules from structured data,resulting in increased efficiency and transparency.Removing the need for human rule construction saves time and money while guaranteeing that users can readily check and trust the automatic judgments of the system.The remarkable performance metrics of the framework,which achieve 99.85%accuracy and 96.30%precision,further support its efficiency in translating complex data into comprehensible rules,eventually empowering users and enhancing organizational decision-making processes.

Measuring Causal Effect with ARDL-BART: A Macroeconomic Application

Modeling dynamic systems with linear parametric models usually suffer limitation which affects forecasting performance and policy implications. This paper advances a non-parametric autoregressive distributed lag model that employs a Bayesian additive regression tree (BART). The performance of the BART model is compared with selection models like Lasso, Elastic Net, and Bayesian networks in simulation experiments with linear and non-linear data generating processes (DGP), and on US macroeconomic time series data. The results show that the BART model is quite competitive against the linear parametric methods when the DGP is linear, and outperforms the competing methods when the DGP is non-linear. The empirical results suggest that the BART estimators are generally more efficient than the traditional linear methods when modeling and forecasting macroeconomic time series.

Spatial pattern of impervious surfaces and their impacts on land surface temperature in Beijing,China

Land surface temperature （LST）, which is heavily influenced by urban surface structures, is a significant parameter in urban environmental analysis. This study examined the effect impervious surfaces （IS） spatial patterns have on LST in Beijing, China. A classification and regression tree model （CART） was adopted to estimate IS as a continuous variable using Landsat images from two seasons combined with QuickBird. LST was retrieved from the Landsat Thematic Mapper （TM） image to examine the relationships between IS and LST. The results revealed that CART was capable of consistently predicting LST with acceptable accuracy （correlation coefficient of 0.94 and the average error of 8.59%）. Spatial patterns of IS exhibited changing gradients across the various urban-rural transects, with LST values showing a concentric shape that increased as you moved from the outskirts towards the downtown areas. Transect analysis also indicated that the changes in both IS and LST patterns were similar at various resolution levels, which suggests a distinct linear relationship between them. Results of correlation analysis further showed that IS tended to be positively correlated with LST, and that the correlation coefficients increased from 0.807 to 0.925 with increases in IS pixel size. The findings identified in this study provide a theoretical basis for improving urban planning efforts to lessen urban temperatures and thus dampen urban heat island effects.

Development of ensemble learning models to evaluate the strength of coal-grout materials

In the loose and fractured coal seam with particularly low uniaxial compressive strength(UCS),driving a roadway is extremely difficult as roof falling and wall spalling occur frequently.To address this issue,the jet grouting(JG)technique(high-pressure grout mixed with coal particles)was first introduced in this study to improve the self-supporting ability of coal mass.To evaluate the strength of the jet-grouted coal-grout composite(JG composite),the UCS evolution patterns were analyzed by preparing 405 specimens combining the influential variables of grout types,curing time,and coal to grout(C/G)ratio.Furthermore,the relationships between UCS and these influencing variables were modeled using ensemble learning methods i.e.gradient boosted regression tree(GBRT)and random forest(RF)with their hyperparameters tuned by the particle swarm optimization(PSO).The results showed that the chemical grout composite has higher short-term strength,while the cement grout composite can achieve more stable strength in the long term.The PSO-GBRT and PSO-RF models can both achieve high prediction accuracy.Also,the variable importance analysis demonstrated that the grout type and curing time should be considered carefully.This study provides a robust intelligent model for predicting UCS of JG composites,which boosts JG design in the field.

Ladle Furnace Temperature Prediction Model Based on Large-scale Data With Random Forest

In ladle furnace, the prediction of the liquid steel temperature is always a hot topic for the researchers. The most of the existing temperature prediction models use small sample set. Today, the precision of them can not satisfy practical production. Fortunately, the large sample set is accumulated from the practical production process. However, a large sample set makes it difficult to build a liquid steel temperature model. To deal with the issue, the random forest method is preferred in this paper, which is a powerful regression method with low complexity and can be designed very quickly. It is with the parallel ensemble structure,uses sample subsets,and employs a simple learning algorithm of sub-models. Then, the random forest method is applied to establish a temperature model by using the data sampled from the production process. The experiments show that the random forest temperature model is more precise than other temperature models.

Optimization of environmental variables in habitat suitability modeling for mantis shrimp Oratosquilla oratoria in the Haizhou Bay and adjacent waters

Habitat suitability index(HSI)models have been widely used to analyze the relationship between species abundance and environmental factors,and ultimately inform management of marine species.The response of species abundance to each environmental variable is different and habitat requirements may change over life history stages and seasons.Therefore,it is necessary to determine the optimal combination of environmental variables in HSI modelling.In this study,generalized additive models(GAMs)were used to determine which environmental variables to be included in the HSI models.Significant variables were retained and weighted in the HSI model according to their relative contribution(%)to the total deviation explained by the boosted regression tree(BRT).The HSI models were applied to evaluate the habitat suitability of mantis shrimp Oratosquilla oratoria in the Haizhou Bay and adjacent areas in 2011 and 2013–2017.Ontogenetic and seasonal variations in HSI models of mantis shrimp were also examined.Among the four models(non-optimized model,BRT informed HSI model,GAM informed HSI model,and both BRT and GAM informed HSI model),both BRT and GAM informed HSI model showed the best performance.Four environmental variables(bottom temperature,depth,distance offshore and sediment type)were selected in the HSI models for four groups(spring-juvenile,spring-adult,falljuvenile and fall-adult)of mantis shrimp.The distribution of habitat suitability showed similar patterns between juveniles and adults,but obvious seasonal variations were observed.This study suggests that the process of optimizing environmental variables in HSI models improves the performance of HSI models,and this optimization strategy could be extended to other marine organisms to enhance the understanding of the habitat suitability of target species.

Spatial variation and prediction of forest biomass in a heterogeneous landscape

Large areas assessments of forest bioinass distribution are a challenge in heterogeneous landscapes, where variations in tree growth and species composition occur over short distances. In this study, we use statistical and geospatial modeling on densely sample.d forest biomass data to analyze the relative importance of ecological and physiographic variables as determinants of spatial variation of forest biomass in the environmentally heterogeneous region of the Big Sur, California. We estimated biomass in 280 forest Plots （one plot per 2.85 km2） and meas- ured an array of ecological （vegetation community type, distance to edge, amount of surrounding non-forest vegetation, soil properties, fire history） and physiographic drivers （elevation, potential soil moisture and solar radiation, proximity to the coast） of tree growth at each plot location. Our geostatistical analyses revealed that biomass distribution is spatially structured and autocorrelated up to 3.1 kin. Regression tree （RT） models showed that both physiographic and ecological factors influenced bio- mass distribution. Across randomly selected sample densities （sample size 112 to 280）, ecological effects of vegetation community type and distance to forest edge, and physiographic effects of elevation, potential soil moisture and solar radiation were the most consistent predictors of biomass. Topographic moisture index and potential solar radiation had apositive effect on biomass, indicating the importance of topographically- mediated energy and moisture on plant growth and biomass accumula- tion. RT model explained 35% of the variation in biomass and spatially autocorrelated variation were retained in regession residuals. Regression kriging model, developed from RT combined with kriging of regression residuals, was used to map biomass across the Big Sur. This study dem- onstrates how statistical and geospatial modeling can be used to dis- criminate the relative importance of physiographic and ecologic effects on forest biomass and develop spatial models to predict and map biomass distribution across a heterogeneous landscape.

Machine learning-based constitutive models for cement-grouted coal specimens under shearing

Cement-based grouting has been widely used in mining engineering;its constitutive law has not been comprehensively studied.In this study,a novel constitutive law of cement-grouted coal specimens(CGCS)was developed using hybrid machine learning(ML)algorithms.Shear tests were performed on CGCS for the analysis of stress-strain curves and the preparation of the dataset.To maintain the interpretation of the trained ML models,regression tree(RT)was used as the main technique.The effect of maximum RT depth(Maxdepth)on its performance was studied,and the hyperparameters of RT were tuned using the genetic algorithm(GA).The RT performance was also compared with ensemble learning techniques.The optimum correlation coefficient on the training set was determined as 0.835,0.946,0.981,and 0.985 for RT models with Maxdepth=3,5,7,and 9,respectively.The overall correlation coefficient was over 0.9 when the Maxdepth≥5,indicating that the constitutive law of CGCS can be well described.However,the failure type of CGCS could not be captured using the trained RT models.Random forest was found to be the optimum algorithm for the constitutive modeling of CGCS,while RT with the Maxdepth=3 performed the worst.