Gradient Boosting Decision Tree Algorithm Research Articles

Predicting future amyloid biomarkers in dementia patients with machine learning to improve clinical trial patient selection.

IntroductionIn Alzheimer's disease, asymptomatic patients may have amyloid deposition, but predicting their progression rate remains a substantial challenge with implications for clinical trial enrollment. Here, we demonstrate an artificial intelligence approach to use baseline clinical information and images to predict changes in quantitative biomarkers of brain pathology on future images.MethodsPatients from the Alzheimer's Disease Neuroimaging Initiative (ADNI) who underwent positron emission tomography (PET) with the amyloid radiotracer 18F‐AV45 (florbetapir) were included. We identified important baseline PET image features using a deep convolutional neural network based on ResNet. These were combined with eight clinical, demographic, and genetic markers using a gradient‐boosted decision tree (GBDT) algorithm to predict future quantitative standardized uptake value ratio (SUVR), an established biomarker of brain amyloid deposition. We used this model to better identify individuals with the highest positive change in amyloid deposition on future images and compared this to typical inclusion criteria for clinical trials. We also compared the model's performance to other methods such as multivariate linear regression and GBDT without imaging features.FindingsUsing 2577 PET scans from 1224 unique individuals, we showed that the GBDT with deep image features was significantly more accurate than the other approaches, reaching a root mean squared error of 0.0339 ± 0.0027 for future SUVR prediction. Using this approach, we could identify individuals with the highest 10% SUVR accumulation at rates 2‐ to 4‐fold higher than by random pick or existing inclusion criteria.DiscussionPredicting quantitative biomarkers on future images using machine learning methods consisting of deep image features combined with clinical data may allow better targeting of treatments or enrollment in clinical trials.

Improving the Resolution and Accuracy of Groundwater Level Anomalies Using the Machine Learning-Based Fusion Model in the North China Plain.

The launch of GRACE satellites has provided a new avenue for studying the terrestrial water storage anomalies (TWSA) with unprecedented accuracy. However, the coarse spatial resolution greatly limits its application in hydrology researches on local scales. To overcome this limitation, this study develops a machine learning-based fusion model to obtain high-resolution (0.25°) groundwater level anomalies (GWLA) by integrating GRACE observations in the North China Plain. Specifically, the fusion model consists of three modules, namely the downscaling module, the data fusion module, and the prediction module, respectively. In terms of the downscaling module, the GRACE-Noah model outperforms traditional data-driven models (multiple linear regression and gradient boosting decision tree (GBDT)) with the correlation coefficient (CC) values from 0.24 to 0.78. With respect to the data fusion module, the groundwater level from 12 monitoring wells is incorporated with climate variables (precipitation, runoff, and evapotranspiration) using the GBDT algorithm, achieving satisfactory performance (mean values: CC: 0.97, RMSE: 1.10 m, and MAE: 0.87 m). By merging the downscaled TWSA and fused groundwater level based on the GBDT algorithm, the prediction module can predict the water level in specified pixels. The predicted groundwater level is validated against 6 in-situ groundwater level data sets in the study area. Compare to the downscaling module, there is a significant improvement in terms of CC metrics, on average, from 0.43 to 0.71. This study provides a feasible and accurate fusion model for downscaling GRACE observations and predicting groundwater level with improved accuracy.

Gradient Boosting Decision Tree Algorithm for the Prediction of Postoperative Intraocular Lens Position in Cataract Surgery.

PurposeTo develop a method for predicting postoperative anterior chamber depth (ACD) in cataract surgery patients based on preoperative biometry, demographics, and intraocular lens (IOL) power.MethodsPatients who underwent cataract surgery and had both preoperative and postoperative biometry measurements were included. Patient demographics and IOL power were collected from the Sight Outcomes Research Collaborative (SOURCE) database. A gradient-boosting decision tree model was developed to predict the postoperative ACD. The mean absolute error (MAE) and median absolute error (MedAE) were used as evaluation metrics. The performance of the proposed method was compared with five existing formulas.ResultsIn total, 847 patients were assigned randomly in a 4:1 ratio to a training/validation set (678 patients) and a testing set (169 patients). Using preoperative biometry and patient sex as predictors, the presented method achieved an MAE of 0.106 ± 0.098 (SD) on the testing set, and a MedAE of 0.082. MAE was significantly lower than that of the five existing methods (P < 0.01). When keratometry was excluded, our method attained an MAE of 0.123 ± 0.109, and a MedAE of 0.093. When IOL power was used as an additional predictor, our method achieved an MAE of 0.105 ± 0.091 and a MedAE of 0.080.ConclusionsThe presented machine learning method achieved greater accuracy than previously reported methods for the prediction of postoperative ACD.Translational RelevanceIncreasing accuracy of postoperative ACD prediction with the presented algorithm has the potential to improve refractive outcomes in cataract surgery.

Development and validation of a machine learning-based postpartum depression prediction model: A nationwide cohort study.

Currently, postpartum depression (PPD) screening is mainly based on self-report symptom-based assessment, with lack of an objective, integrative tool which identifies women at increased risk, before the emergent of PPD. We developed and validated a machine learning-based PPD prediction model utilizing electronic health record (EHR) data, and identified novel PPD predictors. A nationwide longitudinal cohort that included 214,359 births between January 2008 and December 2015, divided into model training and validation sets, was constructed utilizing Israel largest health maintenance organization's EHR-database. PPD was defined as new diagnosis of a depressive episode or antidepressant prescription within the first year postpartum. A gradient-boosted decision tree algorithm was applied to EHR-derived sociodemographic, clinical, and obstetric features. Among the birth cohort, 1.9% (n = 4104) met the case definition of new-onset PPD. In the validation set, the prediction model achieved an area under the curve (AUC) of 0.712 (95% confidence interval, 0.690-0.733), with a sensitivity of 0.349 and a specificity of 0.905 at the 90th percentile risk threshold, identifying PPDs at a rate more than three times higher than the overall set (positive and negative predictive values were 0.074 and 0.985, respectively). The model's strongest predictors included both well-recognized (e.g., past depression) and less-recognized (differing patterns of blood tests) PPD risk factors. Machine learning-based models incorporating EHR-derived predictors, could augment symptom-based screening practice by identifying the high-risk population at greatest need for preventive intervention, before development of PPD.

Машинное обучение как инструмент повышения эффективности управления жизненным циклом высоковольтного электрооборудования

The purpose of the study is to analyze the practical implementation of high-voltage electrical equipment technical state estimation subsystems as a part of solving the lifecycle management problem based on machine learning methods and taking into account the effect of the adjacent power system operation modes. To deal with the problem of power equipment technical state analysis, i.e. power equipment state pattern recognition, XGBoost based on gradient boosting decision tree algorithm is used. Its main advantages are the ability to process gapped data and efficient operation with tabular data for solving classification and regression problems. The author suggests the formation procedure of correct and sufficient initial database for high-voltage equipment state pattern recognition based on its technical diagnostic data and the algorithm for training and testing sets creation in order to improve the identification accuracy of power equipment actual state. The description and justification of the machine learning method and corresponding error metrics are also provided. Based on the actual states of power transformers and circuit breakers the sets of technical diagnostic parameters that have the greatest impact on the accuracy of state identification are formed. The effectiveness of using power systems operation parameters as additional features is also confirmed. It is determined that the consideration of operation parameters obtained by calculation as a part of the training set for high-voltage equipment technical state identification makes it possible to improve the tuning accuracy. The developed structure and approaches to power equipment technical state analysis supplemented by power system operation mode data and diagnostic results provide an information link between the tasks of technological and dispatch control. This allows us to consider the task of power system operation mode planning from the standpoint of power equipment technical state and identify the priorities in repair and maintenance to eliminate power network “bottlenecks”.

Improving the prediction of an atmospheric chemistry transport model using gradient-boosted regression trees

Abstract. Predictions from process-based models of environmental systems are biased, due to uncertainties in their inputs and parameterizations, reducing their utility. We develop a predictor for the bias in tropospheric ozone (O3, a key pollutant) calculated by an atmospheric chemistry transport model (GEOS-Chem), based on outputs from the model and observations of ozone from both the surface (EPA, EMEP, and GAW) and the ozone-sonde networks. We train a gradient-boosted decision tree algorithm (XGBoost) to predict model bias (model divided by observation), with model and observational data for 2010–2015, and then we test the approach using the years 2016–2017. We show that the bias-corrected model performs considerably better than the uncorrected model. The root-mean-square error is reduced from 16.2 to 7.5 ppb, the normalized mean bias is reduced from 0.28 to −0.04, and Pearson's R is increased from 0.48 to 0.84. Comparisons with observations from the NASA ATom flights (which were not included in the training) also show improvements but to a smaller extent, reducing the root-mean-square error (RMSE) from 12.1 to 10.5 ppb, reducing the normalized mean bias (NMB) from 0.08 to 0.06, and increasing Pearson's R from 0.76 to 0.79. We attribute the smaller improvements to the lack of routine observational constraints for much of the remote troposphere. We show that the method is robust to variations in the volume of training data, with approximately a year of data needed to produce useful performance. Data denial experiments (removing observational sites from the algorithm training) show that information from one location (for example Europe) can reduce the model bias over other locations (for example North America) which might provide insights into the processes controlling the model bias. We explore the choice of predictor (bias prediction versus direct prediction) and conclude both may have utility. We conclude that combining machine learning approaches with process-based models may provide a useful tool for improving these models.

Wind Speed Forecast Based on Post-Processing of Numerical Weather Predictions Using a Gradient Boosting Decision Tree Algorithm

With the large-scale development of wind energy, wind power forecasting plays a key role in power dispatching in the electric power grid, as well as in the operation and maintenance of wind farms. The most important technology for wind power forecasting is forecasting wind speed. The current mainstream methods for wind speed forecasting involve the combination of mesoscale numerical meteorological models with a post-processing system. Our work uses the WRF model to obtain the numerical weather forecast and the gradient boosting decision tree (GBDT) algorithm to improve the near-surface wind speed post-processing results of the numerical weather model. We calculate the feature importance of GBDT in order to find out which feature most affects the post-processing wind speed results. The results show that, after using about 300 features at different height and pressure layers, the GBDT algorithm can output more accurate wind speed forecasts than the original WRF results and other post-processing models like decision tree regression (DTR) and multi-layer perceptron regression (MLPR). Using GBDT, the root mean square error (RMSE) of wind speed can be reduced from 2.7–3.5 m/s in the original WRF result by 1–1.5 m/s, which is better than DTR and MLPR. While the index of agreement (IA) can be improved by 0.10–0.20, correlation coefficient be improved by 0.10–0.18, Nash–Sutcliffe efficiency coefficient (NSE) be improved by −0.06–0.6. It also can be found that the feature which most affects the GBDT results is the near-surface wind speed. Other variables, such as forecast month, forecast time, and temperature, also affect the GBDT results.

Mining latent academic social relationships by network fusion of multi-type data

The relationship extraction and network fusion are hotspots of current research in social network mining. Since the types of data are manifold, researchers can utilize multi-type data to construct multiple networks. In academic social network mining, present researches are mostly based on the single-type data, e.g., the co-authorship network constructed by academic co-authorship records. However, the relationships portrayed by single-type data are not sufficient to characterize the complex relationships of the real world. To solve this problem, we are the first to the best of our knowledge to use acknowledgment text to construct a semantic information-based academic social network. First, we extract named entities from multi-type data and implement network optimization and alignment. Second, a semi-supervised fusion framework for multiple networks (SFMN), using the gradient boosting decision tree algorithm to fuse the information of multiple networks into a single network, is proposed in this paper. Third, we implement the parallel version of SFMN on Spark to improve the performance of large-scale social network analysis. Experiments show the superiority of our framework over several state-of-the-art methods and prove our method can effectively integrate network information.

SETE: Sequence-based Ensemble learning approach for TCR Epitope binding prediction

Predicting the binding of T cell receptors (TCRs) to epitopes plays a vital role in the immunotherapy, because it guides the development of therapeutic vaccines and cancer treatments. Many prediction methods attempted to explain the relationship between TCR repertoires from different aspects such as the V(D)J gene locus and the biophysical features of amino acids molecules, but the extraction of these features is time consuming and the performance of these models are limited. Few studies have investigated how k-mers formed by adjacent amino acids in TCR sequences direct the epitope recognition, and the specific mechanism of TCR epitope binding is still unclear. Motivated by these, we presented SETE (Sequence-based Ensemble learning approach for TCR Epitope binding prediction), a novel model to predict the TCR epitope binding accurately. The model deconstructed the CDR3β sequence to short amino acid chains as features and learned the pattern of them between different TCR repertoires with gradient boosting decision tree algorithm. Experiments have demonstrated that SETE can be helpful in predicting the TCRs’ corresponding epitopes and it outperforms other state-of-the-art methods in predicting the epitope specificity of TCR on VDJdb data set. The source codes have been uploaded at https://github.com/wonanut/SETE for academic usage only.

Predicting Hard Rock Pillar Stability Using GBDT, XGBoost, and LightGBM Algorithms

Predicting pillar stability is a vital task in hard rock mines as pillar instability can cause large-scale collapse hazards. However, it is challenging because the pillar stability is affected by many factors. With the accumulation of pillar stability cases, machine learning (ML) has shown great potential to predict pillar stability. This study aims to predict hard rock pillar stability using gradient boosting decision tree (GBDT), extreme gradient boosting (XGBoost), and light gradient boosting machine (LightGBM) algorithms. First, 236 cases with five indicators were collected from seven hard rock mines. Afterwards, the hyperparameters of each model were tuned using a five-fold cross validation (CV) approach. Based on the optimal hyperparameters configuration, prediction models were constructed using training set (70% of the data). Finally, the test set (30% of the data) was adopted to evaluate the performance of each model. The precision, recall, and F1 indexes were utilized to analyze prediction results of each level, and the accuracy and their macro average values were used to assess the overall prediction performance. Based on the sensitivity analysis of indicators, the relative importance of each indicator was obtained. In addition, the safety factor approach and other ML algorithms were adopted as comparisons. The results showed that GBDT, XGBoost, and LightGBM algorithms achieved a better comprehensive performance, and their prediction accuracies were 0.8310, 0.8310, and 0.8169, respectively. The average pillar stress and ratio of pillar width to pillar height had the most important influences on prediction results. The proposed methodology can provide a reliable reference for pillar design and stability risk management.

A gradient boosting decision tree algorithm combining synthetic minority oversampling technique for lithology identification

Lithology identification based on conventional well-logging data is of great importance for geologic features characterization and reservoir quality evaluation in the exploration and production development of petroleum reservoirs. However, there are some limitations in the traditional lithology identification process: (1) It is very time consuming to build a model so that it cannot realize real-time lithology identification during well drilling, (2) it must be modeled by experienced geologists, which consumes a lot of manpower and material resources, and (3) the imbalance of labeled data in well-log data may reduce the classification performance of the model. We have developed a gradient boosting decision tree (GBDT) algorithm combining synthetic minority oversampling technique (SMOTE) to realize fast and automatic lithology identification. First, the raw well-log data are normalized by maximum and minimum normalization algorithm. Then, SMOTE is adopted to balance the number of samples in each class in training process. Next, a lithology identification model is built by GBDT to fit the preprocessed training data set. Finally, the built model is verified with the testing data set. The experimental results indicate that the proposed approach improves the lithology identification performance compared with other machine-learning approaches.

Identification of Tea Red Leaf Spot and Tea Red Scab Based on Hybrid Feature Optimization

Tea leaf diseases seriously affect the quality and the yield of tea. In order to determine whether the tea leaves are infected by diseases or any types of infection, technical support is essential for taking appropriate measures of disease control. Images of normal tea leaves, tea leaves infected with Tea Red Leaf Spot, and leaves infected with Tea Red Scab disease were studied. An identification algorithm for both of the tea leaf diseases based on hybrid feature optimization was proposed. First, the image features were extracted using the Histogram of Oriented Gradient and the Inception v3 model. Then, hybrid feature optimization processing was performed on two types of extracted features. Finally, the Gradient Boosting Decision Tree algorithm was used as the classifier for the identification of tea leaf diseases. Experiments demonstrate that the hybrid feature optimization algorithm reduces the image feature from 36, 068 to less than 150 dimensions while maintaining a high identification accuracy, which greatly reduces the complexity of the identification algorithm. At the same time, the identification accuracy of tea leaf diseases based on hybrid feature optimization algorithm were higher than 95%.

Efficient pyrolysis of ginkgo biloba leaf residue and pharmaceutical sludge (mixture) with high production of clean energy: Process optimization by particle swarm optimization and gradient boosting decision tree algorithm

Production of sustainable clean energy can be achieved by co-pyrolysis of agricultural residues and wastewater sludge. Herein, non-additive thermal behaviour of co-pyrolysis of pharmaceutical sludge and ginkgo biloba leaf residues was investigated. Synergistic effect of co-pyrolysis was not obvious at elevated temperatures. Further, kinetics of co-pyrolysis was studied by fitting Coats-Redfern integration method to thermogravimetric (TG) curve. The change of heat and mass transfer in the reactor caused the change of dynamic parameters. Moreover, hybrid particle swarm optimization and gradient boosting decision tree (PSO–GBDT) algorithm was designed to boost the energy production at full-scale pyrolysis plant by monitoring TG curves. PSO–GBDT model well predicts mass loss rate of the mixture at different heating rates confirming that co-pyrolysis of PS and GBLR can results in high energy production by increasing PS pyrolysis. Designing PSO–GBDT model help to reduced waste production by resourceful treatment of waste in to energy.

Using improved gradient-boosted decision tree algorithm based on Kalman filter (GBDT-KF) in time series prediction

In this study, we analyse two mobile phone activity datasets to predict the future traffic of mobile base stations in urban areas. The predicted time series can be used to reflect the trend of human activity flow. Although common methods such as recurrent neural network and long short-term memory (LSTM) network often achieve a high precision, they have the short back of time-consuming. So, we present the improved gradient-boosted decision tree algorithm based on Kalman filter (GBDT-KF) due to the noise in the original time series, because the decrease in the performance of GBDT is usually caused by overfitting the noise in the signal. According to our experiments, although the RMSE of the predicted values of our GBDT-KF and the ground truth is only 12–14% worse than that of the LSTM model, the proposed GBDT-KF algorithm makes a trade-off between the precision and time complexity and achieves over 100-time training time reduction compared with the LSTM model. By implementing the result of our work, service providers could predict where and when a network congestion would happen; therefore, they could take actions ahead of time. Such applications are useful especially in the era of 5G.

A Performance Fault Diagnosis Method for SaaS Software Based on GBDT Algorithm

SaaS software that provides services through cloud platform has been more widely used nowadays. However, when SaaS software is running, it will suffer from performance fault due to factors such as the software structural design or complex environments. It is a major challenge that how to diagnose software quickly and accurately when the performance fault occurs. For this challenge, we propose a novel performance fault diagnosis method for SaaS software based on GBDT (Gradient Boosting Decision Tree) algorithm. In particular, we leverage the monitoring mean to obtain the performance log and warning log when the SaaS software system runs, and establish the performance fault type set and determine performance log feature. We also perform performance fault type annotation for the performance log combined with the analysis result of the warning log. Moreover, we deal with the incomplete performance log and the type non-equalization problem by using the mean filling for the same type and combination of SMOTE (Synthetic Minority Oversampling Technique) and undersampling methods. Finally, we conduct an empirical study combined with the disaster reduction system deployed on the cloud platform, and it demonstrates that the proposed method has high efficiency and accuracy for the performance diagnosis when SaaS software system runs.

Effective Breast Cancer Recognition Based on Fine-Grained Feature Selection

Early detection and diagnosis of breast cancer are crucial to improve the survival rates of patients. Hence, pathologists and radiologists need a computer-aided diagnosis system to assist their clinical diagnoses effectively and efficiently. However, most breast cancer recognition models are faced with the sample scarcity problem, which results in serious overfitting and lowers recognition performance. To alleviate the sample scarcity problem, a simple, effective model called “refinement, correlation, adaptive” (RCA) for breast cancer recognition is proposed from the perspective of fine-grained feature selection. An innovative multiview efficient range-based gene selection algorithm is proposed to complete the first-layer feature “refinement,” which contributes to suppressing the noisy information in the original feature space. Then, more-discriminant but low-dimensional information among heterogeneous features is mined through the second-layer cross-modal “correlation” mining. Feature dimensions are reduced to a reasonable value that fits the sample size well and alleviates the overfitting problem. Finally, the last-layer decision-tree-guided “adaptive” feature selection is completed using the gradient boosting decision tree algorithm. The RCA model was validated on two well-known datasets. The experimental results demonstrate that the proposed RCA model can address the sample scarcity problem well. It outperforms state-of-the-art baselines, especially in terms of accuracy and the area of the Kiviat diagram. The largest performance improvements of the metrics are 2.39% and 1121, respectively. Moreover, an online diagnosis system based on the RCA model is proposed. It provides rapid and effective breast cancer recognition, which should make clinical diagnoses more convenient and narrow the gap between theoretical research and practical application.

Real-Time Prediction of the Water Accumulation Process of Urban Stormy Accumulation Points Based on Deep Learning

Influenced by climate change and urbanization, urban flood frequently occurs and represents a serious challenge for many cities. Therefore, it is necessary to generate refined predictions of urban floods, such as the prediction of water accumulation processes at water accumulation points, which is of great significance for supporting water-related managers to reduce flood losses. In this study, 16 combination schemes of rainfall sensitivity indicators were used to determine the optimal scheme for predicting the depth of accumulated water, and the gradient boosting decision tree (GBDT) algorithm in deep learning was used to build a prediction model of the accumulation process of urban stormy accumulation points. Among the 16 schemes, the relative error of scheme 1 is 15.39%, and the qualified rate is 92.86%. This scheme exhibits the highest accuracy for the prediction results of water accumulation depth. Given this finding, the GBDT algorithm was used to construct a regression prediction model of the water accumulation process based on the collected historical rainfall water accumulation data of 50 water accumulation points. The results demonstrated that the GBDT regression prediction model has a mean relative error of 19.77%, a qualified rate of 82.00%, and a peak average relative error of 5.48%, which verify the validity and applicability of the model for the real-time prediction of the process of water accumulation.

Exploiting GPUs for Efficient Gradient Boosting Decision Tree Training

In this paper, we present a novel parallel implementation for training Gradient Boosting Decision Trees (GBDTs) on Graphics Processing Units (GPUs). Thanks to the excellent results on classification/regression and the open sourced libraries such as XGBoost, GBDTs have become very popular in recent years and won many awards in machine learning and data mining competitions. Although GPUs have demonstrated their success in accelerating many machine learning applications, it is challenging to develop an efficient GPU-based GBDT algorithm. The key challenges include irregular memory accesses, many sorting operations with small inputs and varying data parallel granularities in tree construction. To tackle these challenges on GPUs, we propose various novel techniques including (i) Run-length Encoding compression and thread/block workload dynamic allocation, (ii) data partitioning based on stable sort, and fast and memory efficient attribute ID lookup in node splitting, (iii) finding approximate split points using two-stage histogram building, (iv) building histograms with the aware of sparsity and exploiting histogram subtraction to reduce histogram building workload, (v) reusing intermediate training results for efficient gradient computation, and (vi) exploiting multiple GPUs to handle larger data sets efficiently. Our experimental results show that our algorithm named ThunderGBM can be 10x times faster than the state-of-the-art libraries (i.e., XGBoost, LightGBM and CatBoost) running on a relatively high-end workstation of 20 CPU cores. In comparison with the libraries on GPUs, ThunderGBM can handle higher dimensional problems which the libraries become extremely slow or simply fail. For the data sets the existing libraries on GPUs can handle, ThunderGBM achieves up to 10 times speedup on the same hardware, which demonstrates the significance of our GPU optimizations. Moreover, the models trained by ThunderGBM are identical to those trained by XGBoost, and have similar quality as those trained by LightGBM and CatBoost.

Incorporating weather conditions and travel history in estimating the alighting bus stops from smart card data

Origin-destination flow of passengers in bus networks is a crucial input to the public transport planning and operational decisions. Smart card systems in many cities, however, record only the bus boarding information (namely an open system), which makes it challenging to use smart card data for origin-destination estimations and subsequent analyses. This study addresses this research gap by proposing a machine learning approach and applying the gradient boosting decision tree (GBDT) algorithm to estimate the alighting stops of bus trips from open smart card data. It advances the state-of-the-art by including, for the first time, weather variables and travel history of individuals in the GBDT algorithm alongside the network characteristics. The method is applied to six-month smart card data from the City of Changsha, China, with more than 17 million trip-records from 700 thousand card users. The model prediction results show that, compared to classic machine learning methods, GBDT not only yields higher prediction accuracy but more importantly is also able to rank the influencing factors on bus ridership. The results demonstrate that incorporation of weather variables and travel history further improves the prediction capability of the models. The proposed GBDT-based framework is flexible and scalable: it can be readily trained with smart card data from other cities to be used for predicting bus origin-destination flow. The results can contribute to improved transport sustainability of a city by enabling smart bus planning and operational decisions.

Schedulable capacity forecasting for electric vehicles based on big data analysis

Fast and accurate forecasting of schedulable capacity of electric vehicles (EVs) plays an important role in enabling the integration of EVs into future smart grids as distributed energy storage systems. Traditional methods are insufficient to deal with large-scale actual schedulable capacity data. This paper proposes forecasting models for schedulable capacity of EVs through the parallel gradient boosting decision tree algorithm and big data analysis for multi-time scales. The time scale of these data analysis comprises the real time of one minute, ultra-short-term of one hour and one-day-ahead scale of 24 hours. The predicted results for different time scales can be used for various ancillary services. The proposed algorithm is validated using operation data of 521 EVs in the field. The results show that compared with other machine learning methods such as the parallel random forest algorithm and parallel k-nearest neighbor algorithm, the proposed algorithm requires less training time with better forecasting accuracy and analytical processing ability in big data environment.