XGBoost Method Research Articles

Identification of phosphorus fractions of biofilm sludge and phosphorus release, transformation and modeling in biofilm sludge treatment related to pH

For seeking out potential phosphorus (P) sources, P fractions of biofilm sludge and P release, transformation and modeling in biofilm sludge treatment related to pH were investigated in this study. A laboratory-scale biofilm system was continuously operated to obtain the biofilm sludge. P release batch experiment was carried out by fermentation at different pH. P fraction and transformation were investigated with sequential extraction scheme and P-31 nuclear magnetic resonance spectroscopy. P release performance model was constructed based on extreme gradient boosting (XGBoost) method. Results showed that inorganic Psolid (IPsolid) occupied most in solid phase of raw biofilm sludge (69.21% total Psolid) and the content of releasable and bioavailable P (part of IPsolid and organic Psolid) was 15.56 mg P/g TSS (90.57% of total Psolid), suggesting that biofilm sludge could be a potential P source. The optimum conditions for P release in biofilm sludge treatment process were pH = 5 and pH = 10, respectively, under acidic and alkaline conditions, and the corresponding maximum concentrations of orthophosphate-P were 72.44 mg/L and 76.84 mg/L. Moreover, pH played an important role on P transformation and redistribution in the treatment process. Credible accuracy of P release performance modeling based on XGBoost method was evaluated by root mean square error, mean absolute error and mean relative error, suggesting the feasibility of XGBoost method in modeling P release process, which was beneficial for the application of P release and recovery from biofilm sludge in practical wastewater treatment process.

Meteorological drought forecasting based on a statistical model with machine learning techniques in Shaanxi province, China

BackgroundDrought is a major natural disaster that causes severe social and economic losses. The prediction of regional droughts may provide important information for drought preparedness and farm irrigation. The existing drought prediction models are mainly based on a single weather station. Efforts need to be taken to develop a new multistation-based prediction model. ObjectivesThis study optimizes the predictor selection process and develops a new model to predict droughts using past drought index, meteorological measures and climate signals from 32 stations during 1961 to 2016 in Shaanxi province, China. MethodsWe applied and compared two methods, including a cross-correlation function and a distributed lag nonlinear model (DLNM), in selecting the optimal predictors and specifying their lag time. Then, we built a DLNM, an artificial neural network model and an XGBoost model and compared their validations for predicting the Standardized Precipitation Evapotranspiration Index (SPEI) 1–6 months in advance. ResultsThe DLNM was better than the cross-correlation function in predictor selection and lag effect determination. The XGBoost model more accurately predicted SPEI with a lead time of 1–6 months than the DLNM and the artificial neural network, with cross-validation R2 values of 0.68–0.82, 0.72–0.89, 0.81–0.92, and 0.84–0.95 at 3-, 6-, 9- and 12-month time scales, respectively. Moreover, the XGBoost model had the highest prediction accuracy for overall droughts (89%–97%) and for three specific drought categories (i.e., moderate, severe, and extreme) (76%–94%). ConclusionThis study offers a new modeling strategy for drought predictions based on multistation data. The incorporation of nonlinear and lag effects of predictors into the XGBoost method can significantly improve prediction accuracy of SPEI and drought.

Street‐segment‐based salt and abrasive prediction for winter maintenance using machine learning and GIS

AbstractSalting operations are essential but expensive in northern winters. We aim to predict the quantity of salt and abrasive needed for a specific road segment for each hour. An estimate of the quantity allows managers to better manage the loads in the trucks and to propose optimal vehicle routes based on the forecast, which will allow them to optimize costs. This article uses machine‐learning techniques based on truck telemetry data, weather conditions, and segment attributes. Geographic information systems (GIS) allow us to exploit the street‐network characteristics, which were ignored by previous prediction models. The results show that the XGBoost method performs better than other techniques (R2 = 0.83).

An interpretable boosting model to predict side effects of analgesics for osteoarthritis

BackgroundOsteoarthritis (OA) is the most common disease of arthritis. Analgesics are widely used in the treat of arthritis, which may increase the risk of cardiovascular diseases by 20% to 50% overall.There are few studies on the side effects of OA medication, especially the risk prediction models on side effects of analgesics. In addition, most prediction models do not provide clinically useful interpretable rules to explain the reasoning process behind their predictions. In order to assist OA patients, we use the eXtreme Gradient Boosting (XGBoost) method to balance the accuracy and interpretability of the prediction model.ResultsIn this study we used the XGBoost model as a classifier, which is a supervised machine learning method and can predict side effects of analgesics for OA patients and identify high-risk features (RFs) of cardiovascular diseases caused by analgesics. The Electronic Medical Records (EMRs), which were derived from public knee OA studies, were used to train the model. The performance of the XGBoost model is superior to four well-known machine learning algorithms and identifies the risk features from the biomedical literature. In addition the model can provide decision support for using analgesics in OA patients.ConclusionCompared with other machine learning methods, we used XGBoost method to predict side effects of analgesics for OA patients from EMRs, and selected the individual informative RFs. The model has good predictability and interpretability, this is valuable for both medical researchers and patients.

Application of XGBoost algorithm in hourly PM2.5 concentration prediction

In view of prediction techniques of hourly PM2.5 concentration in China, this paper applied the XGBoost(Extreme Gradient Boosting) algorithm to predict hourly PM2.5 concentration. The monitoring data of air quality in Tianjin city was analyzed by using XGBoost algorithm. The prediction performance of the XGBoost method is evaluated by comparing observed and predicted PM2.5 concentration using three measures of forecast accuracy. The XGBoost method is also compared with the random forest algorithm, multiple linear regression, decision tree regression and support vector machines for regression models using computational results. The results demonstrate that the XGBoost algorithm outperforms other data mining methods.

A Novel Method of Statistical Line Loss Estimation for Distribution Feeders Based on Feeder Cluster and Modified XGBoost

The estimation of losses of distribution feeders plays a crucial guiding role for the planning, design, and operation of a distribution system. This paper proposes a novel estimation method of statistical line loss of distribution feeders using the feeder cluster technique and modified eXtreme Gradient Boosting (XGBoost) algorithm that is based on the characteristic data of feeders that are collected in the smart power distribution and utilization system. In order to enhance the applicability and accuracy of the estimation model, k-medoids algorithm with weighting distance for clustering distribution feeders is proposed. Meanwhile, a variable selection method for clustering distribution feeders is discussed, considering the correlation and validity of variables. This paper next modifies the XGBoost algorithm by adding a penalty function in consideration of the effect of the theoretical value to the loss function for the estimation of statistical line loss of distribution feeders. The validity of the proposed methodology is verified by 762 distribution feeders in the Shanghai distribution system. The results show that the XGBoost method has higher accuracy than decision tree, neural network, and random forests by comparison of Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), and Absolute Percentage Error (APE) indexes. In particular, the theoretical value can significantly improve the reasonability of estimated results.

Radar emitter classification for large data set based on weighted‐xgboost

Radar emitter classification (REC) is very important in both civil and military fields. It becomes more and more difficult to classify the intercepted radar signals with the increasing complexity of radar signals. An efficient classification method using weighted-xgboost (w-xgboost) model for the complex radar signals is proposed in this study. The xgboost method is widely used by data scientists and performs very well in many machine learning projects. The authors use a large data set which consists of different types of attributes (such as continuous data, categorical data, and discrete data) to train the model. A smooth weight function is introduced to solve the data deviation problem. Experiment results show that the authors' w-xgboost method achieves a better performance than several existing machine learning algorithms on the test set.