Decision Tree-based Models Research Articles

Optimizing inventory control through a data-driven and model-independent framework

Machine learning has shown great potential in various domains, but its appearance in inventory control optimization settings remains rather limited. We propose a novel inventory cost minimization framework that exploits advanced decision-tree based models to approximate inventory performance at an item level, considering demand patterns and key replenishment policy parameters as input. The suggested approach enables data-driven approximations that are faster to perform compared to standard inventory simulations, while being flexible in terms of the methods used for forecasting demand or estimating inventory level, lost sales, and number of orders, among others. Moreover, such approximations can be based on knowledge extracted from different sets of items than the ones being optimized, thus providing more accurate proposals in cases where historical data are scarce or highly affected by stock-outs. The framework was evaluated using part of the M5 competition’s data. Our results suggest that the proposed framework, and especially its transfer learning variant, can result in significant improvements, both in terms of total inventory cost and realized service level.

Detecting Woody Plants in Southern Arizona Using Data from the National Ecological Observatory Network (NEON)

Land cover changes and conversions are occurring rapidly in response to human activities throughout the world. Woody plant encroachment (WPE) is a type of land cover conversion that involves the proliferation and/or densification of woody plants in an ecosystem. WPE is especially prevalent in drylands, where subtle changes in precipitation and disturbance regimes can have dramatic effects on vegetation structure and degrade ecosystem functions and services. Accurately determining the distribution of woody plants in drylands is critical for protecting human and natural resources through woody plant management strategies. Using an object-based approach, we have used novel open-source remote sensing and in situ data from Santa Rita Experimental Range (SRER), National Ecological Observatory Network (NEON), Arizona, USA with machine learning algorithms and tested each model’s efficacy for estimating fractional woody cover (FWC) to quantify woody plant extent. Model performance was compared using standard model assessment metrics such as accuracy, sensitivity, specificity, and runtime to assess model variables and hyperparameters. We found that decision tree-based models with a binary classification scheme performed best, with sequential models (Boosting) slightly outperforming independent models (Random Forest) for both object classification and FWC estimates. Mean canopy height and mean, median, and maximum statistics for all vegetation indices were found to have highest variable importance. Optimal model hyperparameters and potential limitations of the NEON dataset for classifying woody plants in dryland regions were also identified. Overall, this study lays the groundwork for developing machine learning models for dryland woody plant management using solely NEON data.

Ensemble-Based Deep Learning Model for Network Traffic Classification

Network Traffic Classification enables a number of practical applications ranging from network monitoring to resource management, with security implications as well. Nowadays, traffic classification has become a challenging task in order to distinguish among a variety of applications due to the huge amount of generated traffic. Therefore, developing Machine Learning (ML) models, which can successfully identify network applications, is one of the most important tasks. However, among the ML models applied to network traffic classification so far, no model outperforms all the others. To solve these issues, this paper proposes a novel Deep Learning (DL)-based approach that incorporates multiple Decision Tree based models. This approach employs a non-linear blending ensemble method by combining tree-based classifiers through DL in order to maximize generalization accuracy. This ensemble consists of two levels called base classifiers and meta-classifiers. In the first level, Decision Tree-based models are used as the base classifiers while in the second level, DL is used as a meta-model to combine the outputs of the base classifiers. Using two publicly available datasets, we show that our proposed ensemble is suitable for network traffic classification and outperforms the linear blending (using logistic regression as meta-model) as well as several well-known ML models, which are Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), Multi-Layer Perceptron (MLP), AdaBoost, K-Nearest Neighbors (KNN), LightGBM, Catboost, and XGBoost.

Identifying high crash risk segments in rural roads using ensemble decision tree-based models

Traffic safety forecast models are mainly used to rank road segments. While existing studies have primarily focused on identifying segments in urban networks, rural networks have received less attention. However, rural networks seem to have a higher risk of severe crashes. This paper aims to analyse traffic crashes on rural roads to identify the influencing factors on the crash frequency and present a framework to develop a spatial–temporal crash risk map to prioritise high-risk segments on different days. The crash data of Khorasan Razavi province is used in this study. Crash frequency data with the temporal resolution of one day and spatial resolution of 1500 m from loop detectors are analysed. Four groups of influential factors, including traffic parameters (e.g. traffic flow, speed, time headway), road characteristics (e.g. road type, number of lanes), weather data (e.g. daily rainfall, snow depth, temperature), and calendar variables (e.g. day of the week, public holidays, month, year) are used for model calibration. Three different decision tree algorithms, including, Decision Tree (DT), Random Forest (RF) and eXtreme Gradient Boosting (XGBoost) have been employed to predict crash frequency. Results show that based on the traditional evaluation measures, the XGBosst is better for the explanation and interpretation of the factors affecting crash frequency, while the RF model is better for detecting trends and forecasting crash frequency. According to the results, the traffic flow rate, road type, year of the crash, and wind speed are the most influencing variables in predicting crash frequency on rural roads. Forecasting the high and medium risk segment-day in the rural network can be essential to the safety management plan. This risk will be sensitive to real traffic data, weather forecasts and road geometric characteristics. Seventy percent of high and medium risk segment-day are predicted for the case study.

Abstract 13653: Leveraging Natural Language Processing and Machine Learning to Predict Worsening Heart Failure Events

Background: Prior risk models in patients with heart failure (HF) have focused on hospitalizations for worsening HF (WHF) and have not evaluated for differences in predictors by left ventricular ejection fraction (LVEF). We used natural language processing (NLP) and machine learning methods with access to longitudinal electronic health record (EHR) data to develop risk prediction models for WHF events across practice settings and by LVEF category. Methods: We identified all adults with HF and known LVEF on January 1 st of each year from 2011-2019 in an integrated health care system. WHF events within 1 year were defined as any hospitalization, emergency department, or outpatient encounter with ≥1 symptom, ≥2 objective findings including ≥1 sign, and ≥1 change in HF-related therapy. Signs and symptoms were ascertained using rule-based NLP. We conducted boosted decision tree-based ensemble models for any WHF event within each LVEF category: HF with reduced EF (HFrEF; LVEF ≤40%), HF with mildly reduced EF (HFmrEF; LVEF 41-49%), and HF with preserved EF (HFpEF; LVEF ≥50%). We evaluated model discrimination using area under the curve (AUC) and model calibration using Brier scores. Results: Among 359,298 patients from 2011-2019, 65,838 (18%) had HFrEF, 52,491 (15%) had HFmrEF, and 240,969 (67%) had HFpEF. Mean age was 75±12, 47% were women, and 37% were minorities including 10% Black, 11% Asian/Pacific Islander, and 12% of Hispanic ethnicity. WHF events occurred in 22% of patients with HFrEF, 17% with HFmrEF, and 16% with HFpEF. The models displayed an AUC of 0.75 and Brier score of 0.15 for HFrEF and an AUC of 0.77 and Brier scores of 0.12 for both HFmrEF and HFpEF. Clinical predictors were similar across LVEF categories ( Table ). Conclusions: Longitudinal EHR data can be leveraged using NLP and machine learning for accurate risk estimation that reliably identifies clinical predictors across a range of LVEF. These findings may provide novel insight into the natural history of HF.

Multi-Step Ahead Natural Gas Consumption Forecasting Based on a Hybrid Model: Case Studies in The Netherlands and the United Kingdom

With worldwide activities of carbon neutrality, clean energy is playing an important role these days. Natural gas (NG) is one of the most efficient clean energies with less harmful emissions and abundant reservoirs. This work aims at developing a swarm intelligence-based tool for NG forecasting to make more convincing projections of future energy consumption, combining Extreme Gradient Boosting (XGBoost) and the Salp Swarm Algorithm (SSA). The XGBoost is used as the core model in a nonlinear auto-regression procedure to make multi-step ahead forecasting. A cross-validation scheme is adopted to build a nonlinear programming problem for optimizing the most sensitive hyperparameters of the XGBoost, and then the nonlinear optimization is solved by the SSA. Case studies of forecasting the Natural gas consumption (NGC) in the United Kingdom (UK) and Netherlands are presented to illustrate the performance of the proposed hybrid model in comparison with five other intelligence optimization algorithms and two other decision tree-based models (15 hybrid schemes in total) in 6 subcases with different forecasting steps and time lags. The results show that the SSA outperforms the other 5 algorithms in searching the optimal parameters of XGBoost and the hybrid model outperforms all the other 15 hybrid models in all the subcases with average MAPE 4.9828% in NGC forecasting of UK and 9.0547% in NGC forecasting of Netherlands, respectively. Detailed analysis of the performance and properties of the proposed model is also summarized in this work, which indicates it has high potential in NGC forecasting and can be expected to be used in a wider range of applications in the future.

Identification of out-of-hospital cardiac arrest clusters using unsupervised learning

AimOut-of-hospital cardiac arrest (OHCA) is a leading cause of death, and research has identified limitations in analyzing the factors related to the incidence of cardiac arrest and the frequency of bystander cardiopulmonary resuscitation. This study conducts a cluster analysis of the correlation between location-related factors and the outcome of patients with OHCA using two machine learning methods: variational autoencoder (VAE) and the Dirichlet process mixture model (DPMM). MethodsUsing the prospectively collected Smart Advanced Life Support registry in South Korea between August 2015 and December 2018, a secondary retrospective data analysis was performed on patients with OHCA with a presumed cause of cardiac arrest in adults of 18 years or older. VAE and DPMM were used to create clusters to determine groups with a common nature among those with OHCA. ResultsAmong 5876 OHCA cases, 1510 patients were enrolled in the final analysis. Decision tree-based models, which have an accuracy of 95.36%, were also used to interpret the characteristics of clusters. A total of 8 clusters that had similar spatial characteristics were identified using DPMM and VAE. Among the generated clusters, the averages of the four clusters that exhibited a high survival to discharge rate and a favorable neurological outcome were 9.6% and 6.1%, and the averages of the four clusters that exhibited a low outcome were 5.1% and 3.5% respectively. In the decision tree-based models, the most important feature that could affect the prognosis of an OHCA patient was being transferred to a higher-level emergency center. ConclusionThis methodology can facilitate the development of a regionalization strategy that can improve the survival rate of cardiac arrest patients in different regions.

Tree-Based Algorithms and Association Rule Mining for Predicting Patients' Neurological Outcomes After First-Aid Treatment for an Out-of-Hospital Cardiac Arrest During COVID-19 Pandemic: Application of Data Mining.

ObjectiveThe authors performed several tree-based algorithms and an association rules mining as data mining tools to find useful determinants for neurological outcomes in out-of-hospital cardiac arrest (OHCA) patients as well as to assess the effect of the first-aid and basic characteristics in the EMS system.Patients and MethodsThis was a retrospective cohort study. The outcome was Cerebral Performance Categories grading on OHCA patients at hospital discharge. Decision tree-based models inclusive of C4.5 algorithm, classification and regression tree and random forest were built to determine an OHCA patient’s prognosis. Association rules mining was another data mining method which we used to find the combination of prognostic factors linked to the outcome.ResultsThe total of 3520 patients were included in the final analysis. The mean age was 67.53 (±18.4) year-old and 63.4% were men. To overcome the imbalance outcome issue in machine learning, the random forest has a better predictive ability for OHCA patients in overall accuracy (91.19%), weighted precision (88.76%), weighted recall (91.20%) and F1 score (0.9) by oversampling adjustment. Under association rules mining, patients who had any witness on the spot when encountering OHCA or who had ever ROSC during first-aid would be highly correlated with good CPC prognosis.ConclusionThe random forest has a better predictive ability for OHCA patients. This paper provides a role model applying several machine learning algorithms to the first-aid clinical assessment that will be promising combining with Artificial Intelligence for applying to emergency medical services.

Design of predictive model to optimize the solubility of Oxaprozin as nonsteroidal anti-inflammatory drug

These days, many efforts have been made to increase and develop the solubility and bioavailability of novel therapeutic medicines. One of the most believable approaches is the operation of supercritical carbon dioxide fluid (SC-CO2). This operation has been used as a unique method in pharmacology due to the brilliant positive points such as colorless nature, cost-effectives, and environmentally friendly. This research project is aimed to mathematically calculate the solubility of Oxaprozin in SC-CO2 through artificial intelligence. Oxaprozin is a nonsteroidal anti-inflammatory drug which is useful in arthritis disease to improve swelling and pain. Oxaprozin is a type of BCS class II (Biopharmaceutical Classification) drug with low solubility and bioavailability. Here in order to optimize and improve the solubility of Oxaprozin, three ensemble decision tree-based models including random forest (RF), Extremely random trees (ET), and gradient boosting (GB) are considered. 32 data vectors are used for this modeling, moreover, temperature and pressure as inputs, and drug solubility as output. Using the MSE metric, ET, RF, and GB illustrated error rates of 6.29E−09, 9.71E−09, and 3.78E−11. Then, using the R-squared metric, they demonstrated results including 0.999, 0.984, and 0.999, respectively. GB is selected as the best fitted model with the optimal values including 33.15 (K) for the temperature, 380.4 (bar) for the pressure and 0.001242 (mole fraction) as optimized value for the solubility.

Applicability of decision tree-based machine learning models in the prediction of core-calibrated shale facies from wireline logs in the late Devonian Duvernay Formation, Alberta, Canada

Well logs provide insight into stratigraphically compartmentalized rock properties and are a cost-effective alternative to core. The identification of reservoir (and nonreservoir) facies in core, and their calibration to well-log response has traditionally relied on expert domain knowledge and is inherently inconsistent. Such analyses are time-consuming, tedious, error prone, and often biased due to a lack of objectivity. Automated lithologic interpretations from wireline logs appear to be a promising solution for identifying and understanding depositional complexity within a reservoir. Using the Duvernay Formation in the Western Canada Sedimentary Basin as a case study, the authors evaluate the applicability of decision tree-based machine learning (ML) methods in the prediction of core-calibrated facies and/or facies association distributions within wireline logs. The authors use three independent decision tree-based ML models to predict (1) facies (FACM), (2) facies associations (FAM), and (3) reservoir rock (RESM) from wireline logs. Model accuracies are 60.3%, 88.1%, and 88.1% for FACM, FAM, and RESM, respectively, but individual class F1 scores range from 0 to 0.92. The authors attribute discrepancies in individual class performance to interval thickness, sample proportion of training data, and distinguishability of the output class. Classes thicker than 3 m and encompassing at least 16% of the training data set have F1 scores greater than 0.60. The authors attribute exceptions to these general cutoffs to the ability to recognize diagnostic sedimentologic features observed in core. Results from this study help in understanding stratigraphic complexity in the absence of core aiding in subsurface characterization of reservoirs.

Data fusion and machine learning for ship fuel efficiency modeling: Part II – Voyage report data, AIS data and meteorological data

When voyage report data is utilized as the main data source for ship fuel efficiency analysis, its information on weather and sea conditions is often regarded as unreliable. To solve this issue, this study approaches AIS data to obtain the ship's actual detailed geographical positions along its sailing trajectory and then further retrieve the weather and sea condition information from publicly accessible meteorological data sources. These more reliable data about weather and sea conditions the ship sails through is fused into voyage report data in order to improve the accuracy of ship fuel consumption rate models. Eight 8100-TEU to 14,000-TEU containerships from a global shipping company were used in experiments. For each ship, nine datasets were constructed based on data fusion and eleven widely-adopted machine learning models were tested. Experimental results revealed the benefits of fusing voyage report data, AIS data, and meteorological data in improving the fit performances of machine learning models of forecasting ship fuel consumption rate. Over the best datasets, the performances of several decision tree-based models are promising, including Extremely randomized trees (ET), AdaBoost (AB), Gradient Tree Boosting (GB) and XGBoost (XG). With the best datasets, their R2 values over the training sets are all above 0.96 and mostly reach the level of 0.99–1.00, while their R2 values over the test sets are in the range from 0.75 to 0.90. Fit errors of ET, AB, GB, and XG on daily bunker fuel consumption, measured by RMSE and MAE, are usually between 0.8 and 4.5 ton/day. These results are slightly better than our previous study, which confirms the benefits of adopting the actual geographical positions of the ship recorded by AIS data, compared with the estimated geographical positions derived from the great circle route, in retrieving weather and sea conditions the ship sails through.

Comparative study of advanced computational techniques for estimating the compressive strength of UHPC

The effect of raw materials on the compressive strength of concrete is a complex process, especially in the case of ultra-high-performance concrete (UHPC), where a higher number of inter-dependent parameters are involved in the strength development. In this era of digitalization, advanced machine learning methods are used to predict the material's mechanical characteristics because of their superior performance compared to conventional and nonlinear statistical regression models. Thus, the goal of the current study is to estimate the compressive strength of UHPC from the designed raw materials using advanced machine learning techniques. The compressive strength of UHPC is predicted from the 14 input parameters, i.e., cement, fly ash, slag, silica fume, nano-silica, limestone powder, sand, coarse aggregate, quartz powder, water, superplasticizer, PE fiber, steel fiber, and curing time. A total of eight machine learning models were compared that include multi-layer perceptron neural network (MLPNN), MLPNN Bootstrap aggregating (MLPNN-BA), MLPNN adaptive boosting (MLPNN-AB), Gradient boosting (GB), Decision tree (DT), DT Bootstrap aggregating (DT-BA), DT adaptive boosting (DT-AB) and Random Forest (RNF). The validation and performance evaluation of the above models were checked by using K-fold cross-validation, mean absolute error (MAE), root mean square error (RSME), coefficient of determination (R2), relative root mean square error (RRMSE), performance index (PI), and Nash Sutcliffe efficiency (NSE). The optimal model was selected based on the results of all statistical checks. It was found the ensembled machine learning models especially decision tree-based models outperform the neural network-based models with higher accuracy and low error. Thus, the recommended machine learning model is random forest having superior prediction capacity followed by DT Bootstrap aggregating and DT adaptive boosting.

True Reduction in the Air Pollution Levels in the Community of Madrid During the COVID-19 Lockdown

The coronavirus disease (COVID) lockdown was implemented in 2020, which included harsh restrictions on the amount of traffic. As a consequence, a low-emission scenario that could only be simulated before, actually occurred. This constituted a unique and valuable opportunity to study the effect of air quality pollutant concentrations. Although a direct comparison between the observed measured values given by reference air quality stations (AQSs) and values from before the COVID lockdown provides an idea of the pollution reduction, it cannot be separated from the meteorology, and hence, those studies could be misleading. This study used the approach of modeling a normal business day using both air quality and meteorological data from 2017 to 2019 to train machine-learning models to be able to predict what concentration of the three most concerning pollutants (NO2, O3, and PM10) would be given by the meteorological conditions and the time of the year. The XGBoost and LightGBM gradient boosting decision tree-based models were applied to the time series recorded in Madrid and used to predict the expected concentrations in 2020 if no restrictions had been made. The predictions could then be compared to the real observed AQS data to determine the meteorological normalized reductions. The results showed around a 60% reduction in the NO2 at the three types of AQSs (traffic, suburban, and background) during the most restrictive months of the pandemic. The O3 concentration showed a different behavior depending on the type of AQS, pointing to changes in the regime of other pollutants, such as VOCs. The PM10 was the most difficult case to analyze because of its dependence on external transport phenomena, which were difficult to consider in the models. A set of CTM simulations should be done in the future to assess the O3-VOCs-NOx chemistry.

Decision tree-based machine learning models for above-ground biomass estimation using multi-source remote sensing data and object-based image analysis

Forest above-ground biomass (AGB) estimation provides valuable information about the carbon cycle. Thus, the overall goal of this paper is to present an approach to enhance the accuracy of the AGB estimation. The main objectives are to: 1) investigate the performance of remote sensing data sources, including airborne light detection and ranging (LiDAR), optical, SAR, and their combination to improve the AGB predictions, 2) examine the capability of tree-based machine learning models, and 3) compare the performance of pixel-based and object-based image analysis (OBIA). To investigate the performance of machine learning models, multiple tree-based algorithms were fitted to predictors derived from airborne LiDAR data, Landsat, Sentinel-2, Sentinel-1, and PALSAR-2/PALSAR SAR data collected within New York’s Adirondack Park. Combining remote sensing data from multiple sources improved the model accuracy (RMSE: 52.14 Mg ha−1 and R2: 0.49). There was no significant difference among gradient boosting machine (GBM), random forest (RF), and extreme gradient boosting (XGBoost) models. In addition, pixel-based and object-based models were compared using the airborne LiDAR-derived AGB raster as a training/testing sample. The OBIA provided the best results with the RMSE of 33.77 Mg ha−1 and R2 of 0.81 for the combination of optical and SAR data in the GBM model.

A machine learning framework for urban mining: A case study on recovery of copper from printed circuit boards

• Recovery process design by a generic machine-learning-based framework. • Decision tree-based models show similar accuracy; Gradient Boost model outperforms. • Parameter range reduction using LIME and SHAP explanation and interpretation. • Effect of process parameters visualized by multidimensional data plots. Plenty of research articles on developing methods to recover metals from secondary sources have been published. These methods are optimized for a specific source and have poor reproducibility when used for different sources. However, the composition of the source changes with time, manufacturer, and geography, making designing the recovery process a tedious endeavor. A modeling framework that captures the source variation and suggests the process parameters was developed and employed to design a process for copper recovery from the printed circuit board (PCB). Data collected from 23-research articles was visualized using four-dimensional plots. Plots show that the leaching time required for Cu recovery is inversely proportional to hydrogen peroxide concentration, acid concentration, and source % Cu. Recovery is amplified and faster when all these parameters are set to high value, which may not be feasible commercially. Five supervised machine-learning algorithms (support vector machine, random forest, gradient boost machine, XG Boost, and AdaBoost) were trained on 1200 data points as classification and regression problems and validated using a 10-fold cross-validation procedure. Models were tested on 120 data points and compared for predicting accuracy; the gradient boost machine model performs best with an MAE of 10.83% and an F1 score of 0.72. Feature importance analysis based on LIME and permutation importance is used to evaluate the contribution of each feature on recovery, and reduced parameter ranges for high recovery are obtained. Our modeling framework is generic, which can be used for designing any recovery process.

Epileptic seizure classification of electroencephalogram signals using extreme gradient boosting classifier

Epilepsy causes repeated seizures in an individual's life, which causes transient irregularities in the brain's electrical activity. It results in different physical symptoms that are abnormal. Various antiepileptic drugs fail to minimize repeated patient seizures. The electroencephalogram (EEG) signal recordings provide us with time-series data set for epileptic seizure detection and analysis. These signals are highly nonlinear and inconsistent, and they are recorded over time. Predicting the ictal period (seizure period at the time of epilepsy) is thus a challenging task in the naked eye for the medical practitioners. Various machine learning techniques are applied to identify the seizure's occurrence and its classification in multiple domains. A classification model based on extreme gradient boosting (SCLXGB) is proposed here for the classification of the EEG signals. The SCLXGB model implements binary seizure classification on the benchmark dataset. Compared with K-nearest neighbor, linear regression, and Decision treebased models, the proposed model achieves the best area under receiver operating curve (AUC) of 0.9462 and an accuracy of 96% which signifies accurate prediction of seizure and non seizure period. The proposed model SCLXGB was validated by taking different performance metrics to indicate the occurrence and non-occurrence of seizures in patients more appropriately.

Decision Tree-Based Predictive Models for Academic Achievement Using College Students’ Support Networks

In this study, we examine a set of primary data collected from 484 students enrolled in a large public university in the Mid-Atlantic United States region during the early stages of the COVID-19 pandemic. The data, called Ties data, included students’ demographic and support network information. The support network data comprised of information that highlighted the type of support, (i.e. emotional or educational; routine or intense). Using this data set, models for predicting students’ academic achievement, quantified by their self-reported GPA, were created using Chi-Square Automatic Interaction Detection (CHAID), a decision tree algorithm, and cforest, a random forest algorithm that uses conditional inference trees. We compare the methods’ accuracy and variation in the set of important variables suggested by each algorithm. Each algorithm found different variables important for different student demographics with some overlap. For White students, different types of educational support were important in predicting academic achievement, while for non-White students, different types of emotional support were important in predicting academic achievement. The presence of differing types of routine support were important in predicting academic achievement for cisgender women, while differing types of intense support were important in predicting academic achievement for cisgender men.

Modeling of CO2 adsorption capacity by porous metal organic frameworks using advanced decision tree-based models

In recent years, metal organic frameworks (MOFs) have been distinguished as a very promising and efficient group of materials which can be used in carbon capture and storage (CCS) projects. In the present study, the potential ability of modern and powerful decision tree-based methods such as Categorical Boosting (CatBoost), Light Gradient Boosting Machine (LightGBM), Extreme Gradient Boosting (XGBoost), and Random Forest (RF) was investigated to predict carbon dioxide adsorption by 19 different MOFs. Reviewing the literature, a comprehensive databank was gathered including 1191 data points related to the adsorption capacity of different MOFs in various conditions. The inputs of the implemented models were selected as temperature (K), pressure (bar), specific surface area (m2/g) and pore volume (cm3/g) of the MOFs and the output was CO2 uptake capacity (mmol/g). Root mean square error (RMSE) values of 0.5682, 1.5712, 1.0853, and 1.9667 were obtained for XGBoost, CatBoost, LightGBM, and RF models, respectively. The sensitivity analysis showed that among all investigated parameters, only the temperature negatively impacts the CO2 adsorption capacity and the pressure and specific surface area of the MOFs had the most significant effects. Among all implemented models, the XGBoost was found to be the most trustable model. Moreover, this model showed well-fitting with experimental data in comparison with different isotherm models. The accurate prediction of CO2 adsorption capacity by MOFs using the XGBoost approach confirmed that it is capable of handling a wide range of data, cost-efficient and straightforward to apply in environmental applications.

A COMPARISON OF DECISION TREE-BASED MODELS FOR FOREST ABOVE-GROUND BIOMASS ESTIMATION USING A COMBINATION OF AIRBORNE LIDAR AND LANDSAT DATA

Abstract. Forest is one of the most crucial Earth’s resources. Forest above-ground biomass (AGB) mapping has been research endeavors for a long time in many applications since it provides valuable information for carbon cycle monitoring, deforestation, and forest degradation monitoring. A methodology to rapidly and accurately estimate AGB is essential for forest monitoring purposes. Thus, the main objective of this paper was to investigate the performance of decision tree-based models to predict AGB at a site in Huntington Wild Forest (HWF) in Essex County, NY using continuous forest inventory (CFI) plots. The results of decision tree, random forest, and deep forest regression models were compared using light detection and ranging (LiDAR), Landsat 5 TM, and a combination of them. The results illustrated the importance of integration of Landsat 5 TM and LiDAR data, which benefits from both vertical forest structure and spectral information reflected by canopy cover. In addition, the deep forest model with a root mean square error (RMSE) of 51.63 Mg/ha and R-squared (R2) of 0.45 outperformed other regression tree-based models, regardless of the dataset.

Performance of statistical and machine learning-based methods for predicting biogeographical patterns of fungal productivity in forest ecosystems

BackgroundThe prediction of biogeographical patterns from a large number of driving factors with complex interactions, correlations and non-linear dependences require advanced analytical methods and modeling tools. This study compares different statistical and machine learning-based models for predicting fungal productivity biogeographical patterns as a case study for the thorough assessment of the performance of alternative modeling approaches to provide accurate and ecologically-consistent predictions.MethodsWe evaluated and compared the performance of two statistical modeling techniques, namely, generalized linear mixed models and geographically weighted regression, and four techniques based on different machine learning algorithms, namely, random forest, extreme gradient boosting, support vector machine and artificial neural network to predict fungal productivity. Model evaluation was conducted using a systematic methodology combining random, spatial and environmental blocking together with the assessment of the ecological consistency of spatially-explicit model predictions according to scientific knowledge.ResultsFungal productivity predictions were sensitive to the modeling approach and the number of predictors used. Moreover, the importance assigned to different predictors varied between machine learning modeling approaches. Decision tree-based models increased prediction accuracy by more than 10% compared to other machine learning approaches, and by more than 20% compared to statistical models, and resulted in higher ecological consistence of the predicted biogeographical patterns of fungal productivity.ConclusionsDecision tree-based models were the best approach for prediction both in sampling-like environments as well as in extrapolation beyond the spatial and climatic range of the modeling data. In this study, we show that proper variable selection is crucial to create robust models for extrapolation in biophysically differentiated areas. This allows for reducing the dimensions of the ecosystem space described by the predictors of the models, resulting in higher similarity between the modeling data and the environmental conditions over the whole study area. When dealing with spatial-temporal data in the analysis of biogeographical patterns, environmental blocking is postulated as a highly informative technique to be used in cross-validation to assess the prediction error over larger scales.