Search

The uncertainty inherent in three-dimensional (3D) mineral prospectivity mapping (MPM) encompasses (a) mineral system conceptual model uncertainty stemming from geological conceptual frameworks, (b) aleatoric uncertainty, attributable to the variability and noise due to multi-source geoscience datasets collection and processing, as well as 3D geological modeling process, and (c) epistemic uncertainty due to predictive algorithm modeling. Quantifying the uncertainty of 3D MPM is a prerequisite for accepting predictive models in exploration. Previous MPM studies were centered on addressing the mineral system conceptual model uncertainty. To the best of our knowledge, few studies quantified the aleatoric and epistemic uncertainties of 3D MPM. This study proposes a novel uncertainty-quantification machine learning framework to qualify aleatoric and epistemic uncertainties in 3D MPM by the uncertainty-quantification random forest. Another innovation of this framework is utility of the accuracy–rejection curve to provide a quantitative uncertainty threshold for exploration target delineation. The Bayesian hyperparameter optimization tunes the hyperparameters of the uncertainty-quantification random forest automatically. The case study of 3D MPM for exploration target delineation in the Wulong gold district of China demonstrated the practicality of our framework. The aleatoric uncertainty of the 3D MPM indicates that the 3D Early Cretaceous dyke model is the main source of this uncertainty. The 3D exploration targets delineated by the uncertainty-quantification machine learning framework can benefit subsurface gold exploration in the study area. [ABSTRACT FROM AUTHOR]

Tree-based machine learning algorithms, such as random forest, have emerged as effective tools for estimating fine particulate matter (PM2.5) from satellite observations. However, they typically have unchanged model structures and configurations over time and space, and thus may not fully capture the spatiotemporal variations in the relationship between PM2.5 and predictors, resulting in limited accuracy. Here, we propose geographically and temporally weighted tree-based models (GTW-Tree) for remote sensing of surface PM2.5. Unlike traditional tree-based models, GTW-Tree models vary by time and space to simulate the variability in PM2.5 estimation, and they can output variable importance for every location for the deeper understanding of PM2.5 determinants. Experiments in China demonstrate that GTW-Tree models significantly outperform the conventional tree-based models with predictive error reduced by >21%. The GTW-Tree-derived time-location-specific variable importance reveals spatiotemporally varying impacts of predictors on PM2.5. Aerosol optical depth (AOD) contributes largely to PM2.5 estimation, particularly in central China. The proposed models are valuable for spatiotemporal modeling and interpretation of PM2.5 and other various fields of environmental remote sensing. [ABSTRACT FROM AUTHOR]

Landslide susceptibility assessment is an essential tool for disaster prevention and management. In areas with multiple fault zones, the impact of fault zone on slope stability cannot be disregarded. This study performed qualitative analysis of fault zones and proposed a zoning method to assess the landslide susceptibility in Chengkou County, Chongqing Municipality, China. The region within a distance of 1 km from the faults was designated as sub-zone A, while the remaining area was labeled as sub-zone B. To accomplish the assessment, a dataset comprising 388 historical landslides and 388 non-landslide points was used to train the random forest model. 10-fold cross-validation was utilized to select the training and testing datasets for the model. The results of the models were analyzed and discussed, with a focus on model performance and prediction uncertainty. By implementing the proposed division strategy based on fault zone, the accuracy, precision, recall, F-score, and AUC of both two sub-zones surpassed those of the whole region. In comparison to the results obtained for the whole region, sub-zone B exhibited an increase in AUC by 6.15%, while sub-zone A demonstrated a corresponding increase of 1.66%. Moreover, the results of 100 random realizations indicated that the division strategy has little effect on the prediction uncertainty. This study introduces a novel approach to enhance the prediction accuracy of the landslide susceptibility mapping model in areas with multiple fault zones. [ABSTRACT FROM AUTHOR]

Reference crop evapotranspiration (ETO) is a basic component of the hydrological cycle and its estimation is critical for agricultural water resource management and scheduling. In this study, three tree-based machine learning algorithms (random forest [RF], gradient boosting decision tree [GBDT], and extreme gradient boosting [XGBoost]) were adopted to determine the essential factors for ETO prediction. The tree-based models were optimized using the Bayesian optimization (BO) algorithm and were compared with three standalone models in terms of daily ETO and monthly mean ETO estimations in North China, with different input combinations of essential variables. The results indicated that solar radiation (Rs) and air temperature (Ts), including the maximum, minimum, and average temperatures, in daily ETO were the key variables affecting model prediction accuracy. Rs was the most influential factor in the monthly average ETO model followed by Ts. Both relative humidity (RH) and wind speed at 2 m (U2) had little impact on ETO prediction at different scales, although their importance differed. Compared with the GBDT and RF models, the XGBoost model exhibited the best performance for daily ETO and monthly mean ETO estimations. The hybrid tree-based models with the BO algorithm outperformed standalone tree-based models. Overall, compared with the other inputs, the model with three inputs (Rs, Ts, and RH/U2) had the highest accuracy. The BO-XGBoost model exhibited superior performance in terms of the global performance index (GPI) for daily ETO and monthly mean ETO predictions and is recommended as a more accurate model for predicting daily ETO and monthly mean ETO in North China or areas with a similar climate. [ABSTRACT FROM AUTHOR]

Precise and timely prediction of crop yields is crucial for food security and the development of agricultural policies. However, crop yield is influenced by multiple factors within complex growth environments. Previous research has paid relatively little attention to the interference of environmental factors and drought on the growth of winter wheat. Therefore, there is an urgent need for more effective methods to explore the inherent relationship between these factors and crop yield, making precise yield prediction increasingly important. This study was based on four type of indicators including meteorological, crop growth status, environmental, and drought index, from October 2003 to June 2019 in Henan Province as the basic data for predicting winter wheat yield. Using the sparrow search algorithm combined with random forest (SSA-RF) under different input indicators, accuracy of winter wheat yield estimation was calculated. The estimation accuracy of SSA-RF was compared with partial least squares regression (PLSR), extreme gradient boosting (XG-Boost), and random forest (RF) models. Finally, the determined optimal yield estimation method was used to predict winter wheat yield in three typical years. Following are the findings: 1) the SSA-RF demonstrates superior performance in estimating winter wheat yield compared to other algorithms. The best yield estimation method is achieved by four types indicators' composition with SSA-RF) (R2 = 0.805, RRMSE = 9.9%. 2) Crops growth status and environmental indicators play significant roles in wheat yield estimation, accounting for 46% and 22% of the yield importance among all indicators, respectively. 3) Selecting indicators from October to April of the following year yielded the highest accuracy in winter wheat yield estimation, with an R2 of 0.826 and an RAISE of 9.0%. Yield estimates can be completed two months before the winter wheat harvest in June. 4) The predicted performance will be slightly affected by severe drought. Compared with severe drought year (2011) (R2 = 0.680) and normal year (2017) (R2 = 0.790), the SSA-RF model has higher prediction accuracy for wet year (2018) (R2 = 0.820). This study could provide an innovative approach for remote sensing estimation of winter wheat yield, yield. [ABSTRACT FROM AUTHOR]

Topographic variables derived from digital elevation models (DEM) are the most commonly used predictor in soil organic carbon (SOC) prediction. The predictability of these variates for the spatial heterogenity of SOC is influenced by the size of DEM resolution. This paper aimed to investigate the resolution effect of the topography at multiple DEM resolutions and select the most important predictors and the optimal resolution for SOC prediction in a large-scale and morphologically complex area. The study area, covering 15 600 km2, was located in a mountainous province in China. A total of 3901 soil samples and three machine learning algorithms including random forest, gradient boosted regression tree, and artificial neural network were used to construct the SOC predictive models at resolutions of 30, 90, 150, 210, 270, 330, and 390 m. Topographic factors were derived from the DEM with an original resolution of 30 m and were subsequently resampled to varying resolutions by using the bilinear interpolation algorithm. The importance of each variate in each predictive model was computed for the most important predictor selection, and the determination coefficient (R2) and Root Mean Square Error (RMSE) of each predictive model were computed for the accuracy evaluation. Comparative analysis on point-based topographic representation, the accuracy of predictive models, and the variable importance was conducted among different resolutions to investigate the resolution effect on the SOC prediction and to select the optimal resolution for this study area. The results showed that topography was the key factor influencing the SOC distribution in this large-scale area due to the region's mountainous nature, and the relative contribution of the topography to predict SOC distribution varied with DEM resolutions. Elevation was the most important topographic factor at all DEM resolutions. Accuracies of predictive models built on the three machine learning algorithms were all dependent on DEM resolution, indicating that DEM resolution had an important effect on SOC prediction. Overall, the random forest predictive models outperformed the other two algorithms, and its most accurate result was obtained at the 210-m resolution with an R2 of 0.25 and an RMSE of 4.77 g kg–1. In conclusion, we found that finer resolutions did not necessarily produce more accurate SOC predictions even in a large-scale and morphologically complex area. Therefore, investigating the resolution effect is suggested to select an optimal resolution for the SOC prediction in other regions characterized by similar geomorphological conditions as this study area. [ABSTRACT FROM AUTHOR]

Exploring the value of multi-source information fusion to predict small and medium-sized enterprises' (SMEs) credit risk in supply chain finance (SCF) is a popular yet challenging task, as two issues of key variable selection and imbalanced class must be addressed simultaneously. To this end, we develop new forecast models adopting an imbalance sampling strategy based on machine learning techniques and apply these new models to predict credit risk of SMEs in China, using financial information, operation information, innovation information, and negative events as predictors. The empirical results show that the financial-based information, such as TOC, NIR, is most useful in predicting SMEs' credit risk in SCF, and multi-source information fusion is meaningful in better predicting the credit risk. In addition, based on the preferred CSL-RF model, which extends cost-sensitive learning to a random forest, we also present the varying mechanisms of key predictors for SMEs' credit risk by using partial dependency analysis. The strategic insights obtained may be helpful for market participants, such as SMEs' managers, investors, and market regulators. [ABSTRACT FROM AUTHOR]

Temperature-economic growth relationships are computed to quantify the impact of climate change on the economy. However, model performance and differences of predictions among research complicate the use of climate econometric estimation. Machine learning methods provide an alternative that might improve the predictive effects. However, time series and extrapolation issues constrain methods such as random forests. We apply a simple thought experiment with national marginal GDP growth by aggregating subnational climate impact to alleviate the shortcomings in random forests. This paper uses random forests, multivariate cubic regression, and linear spline regression to examine the direct impacts of temperature on economic development and conducts a performance comparison of the methods. The model results indicate an optimal temperature of 15°C, 15°C or 21°C for each model. Furthermore, a thought experiment indicates that the marginal predictions of national GDP changes by approximately 1%, −3%, or −6% for models with 1°C warming. The performance comparison suggests that random forests have stable model performance and better prediction performance in bootstrapping. However, the extrapolation problem in random forests causes underestimation of climate impact in 5% of cells under 6°C warming. Overall, our results suggest that temperature should be considered in economic projections under climate change scenarios. We also suggest the use of more machine learning methods in climate impact assessment. [ABSTRACT FROM AUTHOR]

Mineral prospectivity mapping (MPM) is a fundamental task in mineral exploration. In recent years, the random forest (RF), which is recognized as a significant model for ensemble machine learning algorithms, has been widely used in MPM due to its advantages of high performance and robustness. Nevertheless, the RF method does not fully consider the correlation between variables. In this regard, a hybrid model, namely, the projection pursuit RF (PPRF), has been proposed. In this study, for validation and comparison purposes, the PPRF and ordinary RF methods were employed for the MPM of porphyry Cu–Mo deposits in the Eastern Tianshan orogenic belt, northwestern China, under the same conditions. In the case study, the performance of these two methods was critically investigated. The overall prediction accuracies of the PPRF and RF methods were evaluated by the area under the receiver operating characteristic curve and prediction–area plot approaches. To analyze the sensitivity of the PPRF and RF methods to unbalanced data caused by the limited number of known deposits in the study area, the synthetic minority oversampling technique was used to generate a balanced dataset. The results demonstrated that the PPRF method not only outperforms the ordinary RF method in the overall accuracy of MPM but is also less sensitive to unbalanced data than the RF method. Therefore, the PPRF method provides a novel and useful data-driven method with excellent performance for MPM. It utilizes the combination of predictor variables in the projection feature space rather than the variables themselves to identify the inherent characteristics of a given dataset and may thus discover deep hidden features that the ordinary RF method cannot find. [ABSTRACT FROM AUTHOR]

This study focuses on a traditional business ethics question and aims to use advanced techniques to improve the performance of corporate fraud prediction. Based on the GONE framework, we adopt the machine learning model to predict the occurrence of corporate fraud in China. We first identify a comprehensive set of fraud-related variables and organize them into each category (i.e., Greed, Opportunity, Need, and Exposure) of the GONE framework. Among the six machine learning models tested, the Random Forest (RF) model outperforms the other five models in corporate fraud prediction. Based on the RF model, we show that Exposure variables play a more important role in predicting corporate fraud than other input variables. These results highlight the importance of Exposure variables in corporate fraud prediction and promote the practical use of the machine learning model in solving business ethics questions. [ABSTRACT FROM AUTHOR]

Machine learning is widely used to infer ground-level concentrations of air pollutants from satellite observations. However, a single pollutant is commonly targeted in previous explorations, which would lead to duplication of efforts and ignoration of interactions considering the interactive nature of air pollutants and their common influencing factors. We aim to build a unified model to offer a synchronized estimation of ground-level air pollution levels. We constructed a multi-output random forest (MORF) model and achieved simultaneous estimation of hourly concentrations of PM2.5, PM10, O3, NO2, CO, and SO2 in China, benefiting from the world's first geostationary air-quality monitoring instrument Geostationary Environment Monitoring Spectrometer. MORF yielded a high accuracy with cross-validated R2 reaching 0.94. Meanwhile, model efficiency was significantly improved compared to single-output models. Based on retrieved results, the spatial distributions, seasonality, and diurnal variations of six air pollutants were analyzed and two typical pollution events were tracked. [ABSTRACT FROM AUTHOR]

Purpose: Integrating process-based models with machine learning (ML) is an important way to derive spatially-explicit information on the dynamics of cropland soil organic carbon (SOC) stock. However, the performance of integrating process-based models with different structures remains unclear. The purpose of this study is to develop integrated models by combining multiple process-based models with ML for improving the spatiotemporal dynamics modelling of cropland SOC. Methods: A total of 1219 cropland SOC stock data (0–20 cm, sampled in 1980, 2000, and 2015) were collected from southern Jiangsu Province of China. Three models were built by integrating either Rothamsted carbon model (RothC), microbial-mineral carbon stabilization model (MIMICS), or the both with space-for-time random forest (RF-SFTS). Results: The overall spatiotemporal patterns of SOC dynamics that modelled by RothC, MIMICS, the integration of RothC with RF-SFTS (RF-RothC), the integration of MIMICS with RF-SFTS (RF-MIMICS), and the integration of RothC and MIMICS with RF-SFTS (RF-RothC-MIMICS) were similar, but their prediction accuracies differed significantly. The RF-RothC-MIMICS model had the best performance among the six models, being a lowest root mean squared error (RMSE) of 0.79 kg C m−2 and a highest coefficient of determination (R2) of 0.54, compared to the prediction accuracy of other models. Conclusions: Integrating process-based models differing in structures with ML significantly improved the spatiotemporal modelling of SOC dynamics. Outputs of this study may provide some guidance on more realistic projection of SOC dynamics under future changes in climate, land use, and management practices. [ABSTRACT FROM AUTHOR]

In recent years, the development of machine learning, especially deep learning, has provided new methods and ideas for current climate research. Stacked auto-encoder based on deep learning are used to perform nonlinear down-scaling to compress the degree of freedom of climate variables in the early stage. And climate predictors features are extracted from the summer precipitation in four regions in eastern China, from which key climate predictors affecting summer precipitation in each region are identified. On this basis, a variety of regression methods including random forest in machine learning method were used to construct prediction models for key climate predictors in each region. The best model parameters were determined by the sensitivity test of model parameters to forecast results. Several years of predictions suggest that the method for the forecast of precipitation in eastern China has a very high skill, especially in southern China. The results show that the anomaly consistency of the proposed model in regional prediction is better than that of the model. Compared with the mainstream model, the prediction results in South China can be improved by more than 10%. The method has a good application prospect for summer precipitation prediction in eastern China. [ABSTRACT FROM AUTHOR]

Landslide susceptibility mapping of mountain roads is frequently confronted by insufficient historical landslide sample data, multicollinearity of existing evaluation index factors, and inconsistency of evaluation factors due to regional environmental variations. Then, a single machine learning model can easily become overfitting, thus reducing the accuracy and robustness of the evaluation model. This paper proposes a combined machine-learning model to address the issues. The landslide susceptibility in mountain roads were mapped by using factor analysis to normalize and reduce the dimensionality of the initial condition factor and generating six new combination factors as evaluation indexes. The mountain roads in the Youxi County, Fujian Province, China were used for the landslide susceptibility mapping. Three most frequently used machine learning techniques, support vector machine (SVM), random forest (RF), and artificial neural network (ANN) models, were used to model the landslide susceptibility of the study area and validate the accuracy of this evaluation index system. The global minimum variance portfolio was utilized to construct a machine learning combined model. 5-fold cross-validation, statistical indexes, and AUC (Area Under Curve) values were implemented to evaluate the predictive accuracy of the landslide susceptibility model. The mean AUC values for the SVM, RF, and ANN models in the training stage were 89.2%, 88.5%, and 87.9%, respectively, and 78.0%, 73.7%, and 76.7%, respectively, in the validating stage. In the training and validation stages, the mean AUC values of the combined model were 92.4% and 87.1%, respectively. The combined model provides greater prediction accuracy and model robustness than one single model. [ABSTRACT FROM AUTHOR]

It is difficult to accurately estimate the incidence rate of intimate partner violence (IPV) using traditional social survey methods because IPV victims are often reluctant to disclose their experiences, leading to an underestimation of the incidence rate. To address this issue, we applied machine learning algorithms to predict the incidence rate of IPV in China based on data from the Third Wave Survey on the Social Status of Women in China (TWSSSCW 2010). Specifically, we examined five unbalanced sample-processing methods and six machine learning algorithms, choosing the random under-sampling ensemble method and the random forest algorithm to impute the missing data. Analysis of the complete data showed that the incidence rates of physical violence, verbal violence, and cold violence were 7.10%, 13.74%, and 21.35%, respectively, which were higher than the incidence rates in the original dataset (4.05%, 11.21%, and 17.95%, respectively). The robustness of our findings was further confirmed by analysis using different training sets. Overall, this study demonstrates that better tools need to be developed to accurately estimate the incidence rates of IPV. It also serves as a useful guide for future research that imputes missing data using machine learning. [ABSTRACT FROM AUTHOR]

The identification of multivariate geochemical anomalies is critical in mineral exploration. Machine learning algorithms have been successfully employed in the recognition of multivariate geochemical anomalies in support of mineral exploration, owing to their strong ability to learn the complex relationship between geochemical characteristics and mineralization. However, applications of machine learning algorithms suffer from data redundancy and the curse of dimensionality. In this study, a hybrid model combining t-distributed stochastic neighbor embedding (t-SNE) and random forest (RF) was used to solve the aforementioned problems in geochemical mapping for gold exploration in the northwestern Hubei Province of China. Specifically, t-SNE was used for dimension reduction and feature extraction from the major and trace elements of geochemical survey data, and RF was used for probabilistic classification of geochemical patterns related to gold deposits. A comparative study demonstrated that the hybrid model of t-SNE + RF possesses stronger generalization ability than that of PCA + RF and pure RF. Specifically, after 15 experiments, the mean area under the receiver operator characteristic curve (AUC) values of t-SNE + RF, PCA + RF, and pure RF were 0.83, 0.65, and 0.75, respectively. These results suggest that the hybrid model combining t-SNE and RF can more efficiently recognize geochemical anomalies associated with gold mineralization. Compared with PCA, t-SNE can more effectively identify hidden information in complex and nonlinear geochemical survey data. In addition, it can reduce information redundancy and further improve the efficiency of RF for processing multidimensional geochemical survey data. The high-probability areas obtained by t-SNE + RF showed a strong spatial correlation with known gold deposits, which can provide critical clues for further prospecting in the study area. [ABSTRACT FROM AUTHOR]

Machine learning (ML) has shown high predictive ability in environmental research. Accurate estimation of daily PM2.5 concentrations is a prerequisite to address environmental public health issues. However, studies on the interpretability of ML algorithms were limited. In this study, we aimed to estimate the daily concentrations of PM2.5 at a seasonal level, and to understand the potential mechanisms of ML algorithms' decisions with SHapley Additive exPlanations (SHAP). Daily ground PM2.5 concentrations and meteorological data were obtained from the Beijing Municipal Ecological and Environmental Monitoring Center, and China Meteorological Data Service Centre between December 2013 and 2019 November. We calculated correlation coefficient and variance inflation factor (VIF) to eliminate the variables with collinearity, and recursive feature elimination (RFE) was further used to selected more important predictors. A series of ML algorithms, including linear regression, the variants of linear regression (Ridge, Lasso, Elasticnet), decision tree (DT), k-nearest neighbor (KNN), support vector regression (SVR), ensemble methods (random forest: RF, eXtreme Gradient Boosting: XGBoost), and deep learning (long short-term memory network: LSTM), were developed to estimate seasonal-level daily PM2.5 concentrations. A 10-fold cross validation was used to tune hyperparameters, and root mean square error (RMSE), mean absolute error (MAE), ratio of performance to deviation (RPD), and Lin's concordance correlation coefficient (LCCC) were used to evaluate models' performance. SHAP was performed for local and global interpretability analysis. The results showed that the distribution of PM2.5 concentrations in Beijing showed obvious seasonal patterns. A total of five variables (Precipitation, Mean wind speed, Sunshine duration, Mean surface temperature, Mean relative humidity) were selected for final prediction. LSTM showed much higher accuracy than other traditional ML models, achieved the smallest RMSE of 19.58 µg/m3 and MAE of 15.11 µg/m3. In terms of selected data set, there was acceptable (LCCC = 0.41 ~ 0.52) agreement and accuracy (RPD = 0.97 ~ 1.92) for LSTM. The SHAP analyses revealed that the meteorological factors had different influences in specific predictions, and the complex interactions were also illustrated. These results enhance our understanding of meteorological factors-PM2.5 relationships and explain the mechanisms of ML algorithms' decisions. [ABSTRACT FROM AUTHOR]

In gas drilling operations, the rate of penetration (ROP) parameter has an important influence on drilling costs. Prediction of ROP can optimize the drilling operational parameters and reduce its overall cost. To predict ROP with satisfactory precision, a stacked generalization ensemble model is developed in this paper. Drilling data were collected from a shale gas survey well in Xinjiang, northwestern China. First, Pearson correlation analysis is used for feature selection. Then, a Savitzky-Golay smoothing filter is used to reduce noise in the dataset. In the next stage, we propose a stacked generalization ensemble model that combines six machine learning models: support vector regression (SVR), extremely randomized trees (ET), random forest (RF), gradient boosting machine (GB), light gradient boosting machine (LightGBM) and extreme gradient boosting (XGB). The stacked model generates meta-data from the five models (SVR, ET, RF, GB, LightGBM) to compute ROP predictions using an XGB model. Then, the leave-one-out method is used to verify modeling performance. The performance of the stacked model is better than each single model, with R2 = 0.9568 and root mean square error = 0.4853 m/h achieved on the testing dataset. Hence, the proposed approach will be useful in optimizing gas drilling. Finally, the particle swarm optimization (PSO) algorithm is used to optimize the relevant ROP parameters. [ABSTRACT FROM AUTHOR]

Displacement prediction plays a significant role in the landslide disaster early warning. However, landslide deformation is a complex nonlinear dynamic process, posing difficulties in the displacement prediction especially for the commonly used static models. This study applies an advanced deep machine learning method called gated recurrent unit (GRU) to the displacement prediction of the Jiuxianping landslide, which is a typical reservoir landslide located in the Yunyang County of Chongqing, China. Results show that the GRU-based approach is able to portray the variation of the periodic displacement in the testing dataset with fewer outliers. Although both the artificial neural network (ANN) and random forest regression (RFR) can capture the variation tendency of data points in the training dataset, they are unable to predict the local peaks well in the testing dataset. For the multivariate adaptive regression splines (MARS), the deformation characteristics of the periodic displacement curve cannot be well captured, and the overall predictive performance is unsatisfactory. Different from the three static models, the GRU model is essentially a dynamic model making full use of the historical information, which can portray the deformation characteristics of the Jiuxianping landslide rationally. [ABSTRACT FROM AUTHOR]

Artificial intelligence has been increasingly employed to improve operations for various firms and industries. In this study, we construct a box office revenue prediction system for a film at its early stage of production, which can help management overcome resource allocation challenges considering the significant investment and risk for the whole film production. In this research, we focus on China's film market, the second-largest box office in the world. Our model is based on data regarding the nature of a film itself without word-of-mouth data from social platforms. Combining extreme gradient boosting, random forest, light gradient boosting machine, k-nearest neighbor algorithm, and stacking model fusion theory, we establish a stacking model for film box office prediction. Our empirical results show that the model exhibits good prediction accuracy, with its 1-Away accuracy being 86.46%. Moreover, our results show that star influence has the strongest predictive power in this model. [ABSTRACT FROM AUTHOR]

We focus on the causes of fluctuations in wintertime PM10 in nine regional core cities of China using two machine learning models, Random Forest (RF) and Recurrent Neural Network (RNN). RF and RNN both show high performance in predicting hourly PM10 using only gaseous air pollutants (SO2, NO2 and CO) as inputs, showing the predominance of the secondary inorganic aerosol and implying the existence of thermodynamic equilibrium between gaseous air pollutants and PM10. Also, we find the following results. The correlation of gaseous air pollutants and PM10 were more relevant than that of meteorological conditions and PM10. CO was the predominant factor for PM10 in the Beijing-Tianjin-Hebei Plain and the Yangtze River Delta while SO2 and NO2 were also important features for PM10 in the Pearl River Delta and Sichuan Basin. The spatial heterogeneity and temporal homogeneity of PM10 in China are revealed. The long-range transported PM10 was substantiated to be insignificant, except in the sandstorms. The severity of PM10 was attributable to the lopsided shift of thermodynamic equilibrium and the phenology of indigenous flora. [ABSTRACT FROM AUTHOR]

The early stage lung cancer often appears as ground-glass nodules (GGNs). The diagnosis of GGN as preinvasive lesion (PIL) or invasive adenocarcinoma (IA) is very important for further treatment planning. This paper proposes an automatic GGNs' invasiveness classification algorithm for the adenocarcinoma. 1431 clinical cases and a total of 1624 GGNs (3–30 mm) were collected from Shanghai Cancer Center for the study. The data is in high-resolution computed tomography (HRCT) format. Firstly, the automatic GGN detector which is composed by a 3D U-Net and a 3D multi-receptive field (multi-RF) network detects the location of GGNs. Then, a deep 3D convolutional neural network (3D-CNN) called Attention-v1 is used to identify the GGNs' invasiveness. The attention mechanism was introduced to the 3D-CNN. This paper conducted a contract experiment to compare the performance of Attention-v1, ResNet, and random forest algorithm. ResNet is one of the most advanced convolutional neural network structures. The competition performance metrics (CPM) of automatic GGN detector reached 0.896. The accuracy, sensitivity, specificity, and area under curve (AUC) value of Attention-v1 structure are 85.2%, 83.7%, 86.3%, and 92.6%. The algorithm proposed in this paper outperforms ResNet and random forest in sensitivity, accuracy, and AUC value. The deep 3D-CNN's classification result is better than traditional machine learning method. Attention mechanism improves 3D-CNN's performance compared with the residual block. The automatic GGN detector with the addition of Attention-v1 can be used to construct the GGN invasiveness classification algorithm to help the patients and doctors in treatment. [ABSTRACT FROM AUTHOR]

Coal-floor water-inrush incidents account for a large proportion of coal mine disasters in northern China, and accurate risk assessment is crucial for safe coal production. A novel and promising assessment model for water inrush is proposed based on random forest (RF), which is a powerful intelligent machine-learning algorithm. RF has considerable advantages, including high classification accuracy and the capability to evaluate the importance of variables; in particularly, it is robust in dealing with the complicated and non-linear problems inherent in risk assessment. In this study, the proposed model is applied to Panjiayao Coal Mine, northern China. Eight factors were selected as evaluation indices according to systematic analysis of the geological conditions and a field survey of the study area. Risk assessment maps were generated based on RF, and the probabilistic neural network (PNN) model was also used for risk assessment as a comparison. The results demonstrate that the two methods are consistent in the risk assessment of water inrush at the mine, and RF shows a better performance compared to PNN with an overall accuracy higher by 6.67%. It is concluded that RF is more practicable to assess the water-inrush risk than PNN. The presented method will be helpful in avoiding water inrush and also can be extended to various engineering applications. [ABSTRACT FROM AUTHOR]

Background: Aboveground biomass (AGB) is a fundamental indicator of forest ecosystem productivity and health and hence plays an essential role in evaluating forest carbon reserves and supporting the development of targeted forest management plans. Methods: Here, we proposed a random forest/co-kriging framework that integrates the strengths of machine learning and geostatistical approaches to improve the mapping accuracies of AGB in northern Guangdong Province of China. We used Landsat time-series observations, Advanced Land Observing Satellite (ALOS) Phased Array L-band Synthetic Aperture Radar (PALSAR) data, and National Forest Inventory (NFI) plot measurements, to generate the forest AGB maps at three time points (1992, 2002 and 2010) showing the spatio-temporal dynamics of AGB in the subtropical forests in Guangdong, China. Results: The proposed model was capable of mapping forest AGB using spectral, textural, topographical variables and the radar backscatter coefficients in an effective and reliable manner. The root mean square error of the plot-level AGB validation was between 15.62 and 53.78 t∙ha− 1, the mean absolute error ranged from 6.54 to 32.32 t∙ha− 1, the bias ranged from − 2.14 to 1.07 t∙ha− 1, and the relative improvement over the random forest algorithm was between 3.8% and 17.7%. The largest coefficient of determination (0.81) and the smallest mean absolute error (6.54 t∙ha− 1) were observed in the 1992 AGB map. The spectral saturation effect was minimized by adding the PALSAR data to the modeling variable set in 2010. By adding elevation as a covariable, the co-kriging outperformed the ordinary kriging method for the prediction of the AGB residuals, because co-kriging resulted in better interpolation results in the valleys and plains of the study area. Conclusions: Validation of the three AGB maps with an independent dataset indicated that the random forest/co-kriging performed best for AGB prediction, followed by random forest coupled with ordinary kriging (random forest/ordinary kriging), and the random forest model. The proposed random forest/co-kriging framework provides an accurate and reliable method for AGB mapping in subtropical forest regions with complex topography. The resulting AGB maps are suitable for the targeted development of forest management actions to promote carbon sequestration and sustainable forest management in the context of climate change. [ABSTRACT FROM AUTHOR]