Subspace corrected relevance learning with application in neuroimaging

In machine learning, data often comes from different sources, but combining them can introduce extraneous variation that affects both generalization and interpretability. For example, we investigate the classification of neurodegenerative diseases using FDG-PET data collected from multiple neuroimaging centers. However, data collected at different centers introduces unwanted variation due to differences in scanners, scanning protocols, and processing methods. To address this issue, we propose a two-step approach to limit the influence of center-dependent variation on the classification of healthy controls and early vs. late-stage Parkinson's disease patients. First, we train a Generalized Matrix Learning Vector Quantization (GMLVQ) model on healthy control data to identify a "relevance space" that distinguishes between centers. Second, we use this space to construct a correction matrix that restricts a second GMLVQ system's training on the diagnostic problem. We evaluate the effectiveness of this approach on the real-world multi-center datasets and simulated artificial dataset. Our results demonstrate that the approach produces machine learning systems with reduced bias - being more specific due to eliminating information related to center differences during the training process - and more informative relevance profiles that can be interpreted by medical experts. This method can be adapted to similar problems outside the neuroimaging domain, as long as an appropriate "relevance space" can be identified to construct the correction matrix.

Explore the spatial pattern of carbon emissions in urban functional zones: a case study of Pudong, Shanghai, China

It is crucial for the development of carbon reduction strategies to accurately examine the spatial distribution of carbon emissions. Limited by data availability and lack of industry segmentation, previous studies attempting to model spatial carbon emissions still suffer from significant uncertainty. Taking Pudong New Area as an example, with the help of multi-source data, this paper proposed a research framework for the amount calculation and spatial distribution simulation of its CO2 emissions at the scale of urban functional zones (UFZs). The methods used in this study were based on mapping relations among the locations of geographic entities and data of multiple sources, using the coefficient method recommended by the Intergovernmental Panel on Climate Change (IPCC) to calculate emissions. The results showed that the emission intensity of industrial zones and transport zones was much higher than that of other UFZs. In addition, Moran's I test indicated that there was a positive spatial autocorrelation in high emission zones, especially located in industrial zones. The spatial analysis of CO2 emissions at the UFZ scale deepened the consideration of spatial heterogeneity, which could contribute to the management of low carbon city and the optimal implementation of energy allocation.

Recent trends of machine learning applied to multi-source data of medicinal plants

In traditional medicine and ethnomedicine, medicinal plants have long been recognized as the basis for materials in therapeutic applications worldwide. In particular, the remarkable curative effect of traditional Chinese medicine during corona virus disease 2019 (COVID-19) pandemic has attracted extensive attention globally. Medicinal plants have, therefore, become increasingly popular among the public. However, with increasing demand for and profit with medicinal plants, commercial fraudulent events such as adulteration or counterfeits sometimes occur, which poses a serious threat to the clinical outcomes and interests of consumers. With rapid advances in artificial intelligence, machine learning can be used to mine information on various medicinal plants to establish an ideal resource database. We herein present a review that mainly introduces common machine learning algorithms and discusses their application in multi-source data analysis of medicinal plants. The combination of machine learning algorithms and multi-source data analysis facilitates a comprehensive analysis and aids in the effective evaluation of the quality of medicinal plants. The findings of this review provide new possibilities for promoting the development and utilization of medicinal plants.

Impact of built environment on urban surface temperature based on multi-source data at the community level in Beilin District, Xi'an, China

With the global warming and rapid urbanization in China, the urban built environment has undergone rapid changes, and the land surface temperatures (LSTs) of urban communities have obvious spatial heterogeneity. To explore the key driving factors of community LSTs, the multi-source data and spatial statistical methods being jointly used to analyze the spatial characteristics and main influencing factors of LST at the community level in the Beilin District of Xi'an City, China. The results are as follows: (1) Compared with communities dominated by construction land, communities with large area of green space and water bodies have lower LST. (2) According to the Akaike's information criterion (AICc) and maximum of adjusted R2, and other parameters, the No.1236 model was selected as the optimal model to analyze the influencing factors of community LST by exploratory data analysis, including building density (BD), building height standard deviation (BHS), percentage of public administration and public services land (PASL), percentage of green space and square land (PGSL), population density (POPD), normalized difference impervious surface index (NDISI), and perimeter-area fractal dimension (PAFRAC). (3) For each increase of one unit in NDISI and BHS when other factors remain unchanged, the LST will increase by 0.569 °C and decrease by 0.478 °C, respectively. (4) From the spatial stability and distribution of Local-R2, the warming factors of community LST are mainly NDISI, PAFRAC, BD, and PASL, while the cooling factors are BHS and PGSL. The spatial heterogeneity of community LST is not only related to the change of underlying surface properties but is also affected by intra-urban architectural morphology. Therefore, reasonable planning of urban built environment is of great significance for mitigating heat islands.

Machine learning applications for multi-source data of edible crops: A review of current trends and future prospects

The quality and safety of edible crops are key links inseparable from human health and nutrition. In the era of rapid development of artificial intelligence, using it to mine multi-source information on edible crops provides new opportunities for industrial development and market supervision of edible crops. This review comprehensively summarized the applications of multi-source data combined with machine learning in the quality evaluation of edible crops. Multi-source data can provide more comprehensive and rich information from a single data source, as it can integrate different data information. Supervised and unsupervised machine learning is applied to data analysis to achieve different requirements for the quality evaluation of edible crops. Emphasized the advantages and disadvantages of techniques and analysis methods, the problems that need to be overcome, and promising development directions were proposed. To monitor the market in real-time, the quality evaluation methods of edible crops must be innovated.

A generalizable and interpretable model for mortality risk stratification of sepsis patients in intensive care unit

PURPOSE

This study aimed to construct a mortality model for the risk stratification of intensive care unit (ICU) patients with sepsis by applying a machine learning algorithm.

METHODS

Adult patients who were diagnosed with sepsis during admission to ICU were extracted from MIMIC-III, MIMIC-IV, eICU, and Zigong databases. MIMIC-III was used for model development and internal validation. The other three databases were used for external validation. Our proposed model was developed based on the Extreme Gradient Boosting (XGBoost) algorithm. The generalizability, discrimination, and validation of our model were evaluated. The Shapley Additive Explanation values were used to interpret our model and analyze the contribution of individual features.

RESULTS

A total of 16,741, 15,532, 22,617, and 1,198 sepsis patients were extracted from the MIMIC-III, MIMIC-IV, eICU, and Zigong databases, respectively. The proposed model had an area under the receiver operating characteristic curve (AUROC) of 0.84 in the internal validation, which outperformed all the traditional scoring systems. In the external validations, the AUROC was 0.87 in the MIMIC-IV database, better than all the traditional scoring systems; the AUROC was 0.83 in the eICU database, higher than the Simplified Acute Physiology Score III and Sequential Organ Failure Assessment (SOFA),equal to 0.83 of the Acute Physiology and Chronic Health Evaluation IV (APACHE-IV), and the AUROC was 0.68 in the Zigong database, higher than those from the systemic inflammatory response syndrome and SOFA. Furthermore, the proposed model showed the best discriminatory and calibrated capabilities and had the best net benefit in each validation.

CONCLUSIONS

The proposed algorithm based on XGBoost and SHAP-value feature selection had high performance in predicting the mortality of sepsis patients within 24 h of ICU admission.

Spatial scale analysis for the relationships between the built environment and cardiovascular disease based on multi-source data

To examine what built environment characteristics improve the health outcomes of human beings is always a hot issue. While a growing literature has analyzed the link between the built environment and health, few studies have investigated this relationship across different spatial scales. In this study, eighteen variables were selected from multi-source data and reduced to eight built environment attributes using principal component analysis. These attributes included socioeconomic deprivation, urban density, street walkability, land-use diversity, blue-green space, transportation convenience, ageing, and street insecurity. Multiscale geographically weighted regression was then employed to clarify how these attributes relate to cardiovascular disease at different scales. The results indicated that: (1) multiscale geographically weighted regression showed a better fit of the association between the built environment and cardiovascular diseases than other models (e.g., ordinary least squares and geographically weighted regression), and is thus an effective approach for multiscale analysis of the built environment and health associations; (2) built environment variables related to cardiovascular diseases can be divided into global variables with large scales (e.g., socioeconomic deprivation, street walkability, land-use diversity, blue-green space, transportation convenience, and ageing) and local variables with small scales (e.g., urban density and street insecurity); and (3) at specific spatial scales, global variables had trivial spatial variation across the area, while local variables showed significant gradients. These findings provide greater insight into the association between the built environment and lifestyle-related diseases in densely populated cities, emphasizing the significance of hierarchical and place-specific policy formation in health interventions.

National resilience assessment and improvement based on multi-source data: Evidence from countries along the belt and road

National resilience is a consensus benchmark to characterize the ability of disaster resistance of a country. The occurrence of various disasters and the ravages of COVID-19 have created urgent needs in assessing and improving the national resilience of countries, especially for countries along the Belt and Road (i.e., B&R countries) with multiple disasters with high frequency and great losses. To accurately depict the national resilience profile, a three-dimensional assessment model based on multi-source data is proposed, where the diversity of losses, fusion utilization of disaster and macro-indicator data, and several refined elements are involved. Using the proposed assessment model, the national resilience of 64 B&R countries is clarified based on more than 13,000 records involving 17 types of disasters and 5 macro-indicators. However, their assessment results are not optimistic, the dimensional resilience are generally trend-synchronized and individual difference in a single dimension, and approximately one-half of countries do not obtain resilience growth over time. To further explore the applicable solutions for national resilience improvement, a coefficient-adjusted stepwise regression model with 20 macro-indicator regressors is developed based on more than 19,000 records. This study provides the quantified model support and a solution reference for national resilience assessment and improvement, which contributes to addressing the global national resilience deficit and promoting the high-quality development of B&R construction.

Spatial distribution and suitability evaluation of nighttime tourism in Kunming utilizing multi-source data

In the post-pandemic era, nighttime tourism is vital in promoting the diversified development of tourism, enhancing the vitality of cities, and improving reemployment rates. Using Kunming City, China, as an example, this study used multi-theory and multi-source data to construct an evaluation model of the spatial distribution and suitability of nighttime tourism. The projection pursuit model and spatial analysis method were used to reveal spatial distribution and explore the suitability characteristics and spatial differences of nighttime tourism development. Our results revealed the following: (1) nighttime tourism resources showed a 'large aggregation, small dispersion' spatial distribution pattern, which is characterized by a distribution along the railway line; (2) at present, the spatial distribution of nighttime tourism in Kunming displays a pattern of 'taking a high-density large gathering area as the center and extending around; ' and (3) the most suitable nighttime tourism area accounts for 11.83% of the total land area of Kunming; highly suitable area accounts for 17.53%. The general suitable and unsuitable areas accounted for 43.29% and 27.35%, respectively. The results of this study assist in providing a scientific basis for the strategic planning and development of the nighttime tourism industry in Kunming.

Improving the spatial and temporal estimation of ecosystem respiration using multi-source data and machine learning methods in a rainfed winter wheat cropland

The investigation of ecosystem respiration (RE) and its vital influential factors along with the timely and accurate detection of spatiotemporal variations in RE are essential for guiding agricultural production planning. RE observation in the plot region is primarily based on the laborious chamber method. However, upscaling the spatial-temporal estimates of RE at the canopy scale is still challenging. The present study conducted a field experiment to determine RE using the chamber method. A multi-rotor unmanned aerial vehicle (UAV) equipped with a multispectral camera was employed to acquire the canopy spectral data of wheat during each RE test experiment. Moreover, the agronomic indicators of aboveground plant biomass, leaf area index, leaf dry mass as well as agrometeorological and soil data were measured simultaneously. The study analyzed the potential of multi-information for estimating RE at the field scale and proposed two strategies for RE estimation. In addition, a semiempirical, yet Lloyd and Taylor-based, remote sensing model (LT1-NIR_V) was developed for estimating RE observed across different growth stages with a small margin of error (coefficient of determination [R²] = 0.60-0.64, root-mean-square error [RMSE] = 285.98-316.19 mg m^-2 h^-1). Further, five machine learning (ML) algorithms were utilized to independently estimate RE using two different datasets. The rigorous analyses, which included statistical comparison and cross-validation for estimating RE, confirmed that the XGBoost model, with the highest R² and lowest RMSE (R² = 0.88 and RMSE = 172.70 mg m^-2 h^-1), performed the best among the evaluated ML models. The LT1-NIR_V model was less effective in estimating RE compared with the other ML models. Based on this comprehensive comparison analysis, the ML model can successfully estimate variations in wheat field RE using high-resolution UAV multispectral images and environmental factors from the wheat cropland system, thereby providing a valuable reference for monitoring and upscaling RE observations.

Identification of factors controlling heavy metals/metalloid distribution in agricultural soils using multi-source data

Understanding the factors that controlling the agricultural soil heavy metals/metalloids distribution is vital for cropland soil remediation and management. For this objective, 227 agricultural soils were sampled in the Guanzhong Plain, China, to measure the concentration of five heavy metals (Pb, Cd, Ni, Zn, and Cu) and one metalloid (As) by X-ray fluorescence spectrometer, meanwhile, 24 possible influencing factors to agricultural soil heavy metals/metalloid distribution were collected and grouped into three categories. A sequential multivariate statistical analysis was carried out to provide insight into the controlling factors of soil heavy metals/metalloid distribution, then stepwise multiple linear regression (SMLR) and partial least squares regression (PLS) were used to predict heavy metals/metalloid concentrations in agricultural soil based on the result of soil heavy metals/metalloid controlling factors identification. The results demonstrated the types of soil and land use did not have a substantial effect on soil heavy metals/metalloid distribution, except Zn and Cu. The soil properties category played a major role in influencing the soil heavy metals/metalloid concentration. The concentrations of Mn and Fe, which are the main constitute elements of soil inorganic colloid, were the most significant factors, followed by the concentrations of P, K and Ca. Soil pH and soil organic matter (SOM) content, which are often considered as important factors for soil heavy metals/metalloid distribution, were not important in the present study. The SMLR was more effective than the PLS for predicting soil heavy metals/metalloid content. The results of this study enlighten that future soil heavy metals/metalloid contamination treatment in regions with high pH and low SOM content should concentrate on inorganic colloid particles, which have strong adsorption capacity for soil heavy metals/metalloid and are environmentally friendly. Moreover, the combination of successive multivariate statistical analysis and SMLR provide an effective tool to predict and monitor agricultural soil heavy metals/metalloid distribution, and facilitate the improvement of environmental and territorial management.

sJIVE: Supervised Joint and Individual Variation Explained

Analyzing multi-source data, which are multiple views of data on the same subjects, has become increasingly common in molecular biomedical research. Recent methods have sought to uncover underlying structure and relationships within and/or between the data sources, and other methods have sought to build a predictive model for an outcome using all sources. However, existing methods that do both are presently limited because they either (1) only consider data structure shared by all datasets while ignoring structures unique to each source, or (2) they extract underlying structures first without consideration to the outcome. The proposed method, supervised joint and individual variation explained (sJIVE), can simultaneously (1) identify shared (joint) and source-specific (individual) underlying structure and (2) build a linear prediction model for an outcome using these structures. These two components are weighted to compromise between explaining variation in the multi-source data and in the outcome. Simulations show sJIVE to outperform existing methods when large amounts of noise are present in the multi-source data. An application to data from the COPDGene study explores gene expression and proteomic patterns associated with lung function.

MSPCD: predicting circRNA-disease associations via integrating multi-source data and hierarchical neural network

BACKGROUND

Increasing evidence shows that circRNA plays an essential regulatory role in diseases through interactions with disease-related miRNAs. Identifying circRNA-disease associations is of great significance to precise diagnosis and treatment of diseases. However, the traditional biological experiment is usually time-consuming and expensive. Hence, it is necessary to develop a computational framework to infer unknown associations between circRNA and disease.

RESULTS

In this work, we propose an efficient framework called MSPCD to infer unknown circRNA-disease associations. To obtain circRNA similarity and disease similarity accurately, MSPCD first integrates more biological information such as circRNA-miRNA associations, circRNA-gene ontology associations, then extracts circRNA and disease high-order features by the neural network. Finally, MSPCD employs DNN to predict unknown circRNA-disease associations.

CONCLUSIONS

Experiment results show that MSPCD achieves a significantly more accurate performance compared with previous state-of-the-art methods on the circFunBase dataset. The case study also demonstrates that MSPCD is a promising tool that can effectively infer unknown circRNA-disease associations.

Possibility of the COVID-19 third wave in India: mapping from second wave to third wave

After a consistent drop in daily new coronavirus cases during the second wave of COVID-19 in India, there is speculation about the possibility of a future third wave of the virus. The pandemic is returning in different waves; therefore, it is necessary to determine the factors or conditions at the initial stage under which a severe third wave could occur. Therefore, first, we examine the effect of related multi-source data, including social mobility patterns, meteorological indicators, and air pollutants, on the COVID-19 cases during the initial phase of the second wave so as to predict the plausibility of the third wave. Next, based on the multi-source data, we proposed a simple short-term fixed-effect multiple regression model to predict daily confirmed cases. The study area findings suggest that the coronavirus dissemination can be well explained by social mobility. Furthermore, compared with benchmark models, the proposed model improves prediction R 2 by 33.6%, 10.8%, 27.4%, and 19.8% for Maharashtra, Kerala, Karnataka, and Tamil Nadu, respectively. Thus, the simplicity and interpretability of the model are a meaningful contribution to determining the possibility of upcoming waves and direct pandemic prevention and control decisions at a local level in India.

Analysis of the reasons for the outbreak of Yellow Sea green tide in 2021 based on long-term multi-source data

The green tide disaster is the result of human activities changing the natural environment conditions. What changes have occurred in the environmental factors that affect the green tide outbreak over a long period, and what is the impact of this change on the green tide outbreak? To further understand the outbreak mechanism of green tide, in this study, we used the Google Earth Engine (GEE) platform to extract and analyze the green tide from 2007 to 2021, analyze the long-term trend of various influencing factors (sea surface temperature (SST), sea surface salinity (SSS), photosynthetically available radiation (PAR), precipitation, eutrophication, "nori" aquaculture) in the past 30 years, and explore the impact of each factor on the outbreak of green tide. We found that: 1) SST, seawater eutrophication, and "nori" aquaculture worked together to promote the large-scale outbreak of green tide in 2007; 2) In the context of eutrophication is not effectively controlled, elevated SST, SSS, and PAR will be more conducive to the germination of green tide algae and promote green tide to form a floating state on the sea surface earlier, after that, once there is a year with abundant precipitation, the green tide will break out on a large scale, which is exactly the case in 2021. Exploring the environmental conditions and the long-term regularity of green tide outbreaks to provide a basis for scientific and rational control of green tides.

A graph neural network-based stock forecasting method utilizing multi-source heterogeneous data fusion

The study of the prediction of stock market volatility is of great significance to rationally control financial market risks and increase excessive investment returns and has received extensive attention from academic and commercial circles. However, as a dynamic and complex system, the stock market is affected by multiple factors and has a comprehensive capability to include complex financial data. Given that the explanatory variables of influencing factors are diverse, heterogeneous and complex, the existing intelligent algorithms have great limitations for the analysis and processing of multi-source heterogeneous data in the stock market. Therefore, this study adopts the edge weight and information transmission mechanism suitable for subgraph data to complete node screening, the gate recurrent unit (GRU) and long short-term memory (LSTM) to aggregate subgraph nodes. The compiled data contain the metapaths of three types of index data, and the introduction of the association relationship attention dimension effectively mines the implicit meanings of multi-source heterogeneous data. The metapath attention mechanism is combined with a graph neural network to complete the classification of multi-source heterogeneous graph data, by which the prediction of stock market volatility is realized. The results show that the above method is feasible for the fusion of heterogeneous stock market data and the mining of implicit semantic information of association relations. The accuracy of the proposed method for the prediction of stock market volatility in this study is 16.64% higher than that of the dimensional reduction index and 14.48% higher than that of other methods for the fusion and prediction of heterogeneous data using the same model.

Lessons drawn from China and South Korea for managing COVID-19 epidemic: Insights from a comparative modeling study

We conducted a comparative study of the COVID-19 epidemic in three different settings: mainland China, the Guangdong province of China and South Korea, by formulating two disease transmission dynamics models which incorporate epidemic characteristics and setting-specific interventions, and fitting the models to multi-source data to identify initial and effective reproduction numbers and evaluate effectiveness of interventions. We estimated the initial basic reproduction number for South Korea, the Guangdong province and mainland China as 2.6 (95% confidence interval (CI): (2.5, 2.7)), 3.0 (95%CI: (2.6, 3.3)) and 3.8 (95%CI: (3.5,4.2)), respectively, given a serial interval with mean of 5 days with standard deviation of 3 days. We found that the effective reproduction number for the Guangdong province and mainland China has fallen below the threshold 1 since February 8th and 18th respectively, while the effective reproduction number for South Korea remains high until March 2nd Moreover our model-based analysis shows that the COVID-19 epidemics in South Korean is almost under control with the cumulative confirmed cases tending to be stable as of April 14th. Through sensitivity analysis, we show that a coherent and integrated approach with stringent public health interventions is the key to the success of containing the epidemic in China and especially its provinces outside its epicenter. In comparison, we find that the extremely high detection rate is the key factor determining the success in controlling the COVID-19 epidemics in South Korea. The experience of outbreak control in mainland China and South Korea should be a guiding reference for the rest of the world.

Spatial distribution and influencing factors of litter in urban areas based on machine learning - A case study of Beijing

Littering in urban areas negatively affects their appearance, is harmful to the environment and increases pollution. It is a typical urban problem looming large upon Beijing and other megacities striving for liveability and harmony in economy, society and environment. This study analyzed the amount and spatial distribution of urban litter generation in Beijing based on the Kernel Density Estimation method and Anselin's Local Moran I method. We analyzed multiple factors affecting littering in urban areas based on the random forest machine learning method. The results show that the density distribution of litter presents a typical core edge diffusion spatial distribution pattern. High clusters of litter were found in most regions of Dongcheng District and central regions of Haidian District. We have verified that littering in urban areas is mostly affected by population, POIs (interest points), road networks, and the management of the city environment. Among these, permanent population, level of road cleaning, the presence of branch roads and commercial places are the four most important influencing factors. This study is of great significance to the prevention and treatment of littering in urban areas and can help city managers better address this problem.

Environmental inequity in hilly neighborhood using multi-source data from a health promotion view

The study focused on the justice of residents' opportunity to engage in healthy behavior under different environments is not vast, especially in a hilly dwelling environment. Therefore, this paper investigates environmental inequalities in a hilly urban environment in the context of the booming real estate market in China, comprised of health promotion-related elements, namely, built environment, physical activity facilities, street infrastructure, green spaces, and environmental perceptions. The multi-source data are used to calculate environmental attributes and the socioeconomic status of communities. We take the central districts of Dalian city as the research area and measure environmental equity across different socioeconomic residential areas using the Kruskal-Wallis one-way analysis of variance. The results reveal the spatial disparities in physical activity facilities, street greening, and positive perceptions between different communities. However, green injustice is mitigated in the hilly neighborhoods when we consider only ground-level greenness. This paper studies environmental justice by taking a health-enhancing view, and the results of this study can provide guidance on hilly urban development for government leaders and planners.

The effect of farmland on the surface water of the Aral Sea Region using Multi-source Satellite Data

BACKGROUND

The improper land utilization has brought tremendous pressure on the surface water of the Aral Sea Region in the past decades. It was seriously hindered for construction of the Green Silk Road Economic Belt by the fragile environment. Therefore, it is of great necessity for environmental protection and social development to monitor the change of surface water in the Aral Sea Region.

METHODS

In this study, LandTrendr algorithm was used on Landsat time-series data to characterize the change in farmland on the Google Earth Engine platform. Based on multi-source data, the water area changes of the Aral Sea were extracted based on the Google Earth Engine, and the mean method was utilized to extract the changes in water level and water storage. Finally, a water-farmland coupling degree model was utilized to evaluate the impact of farmland changes on the surface water in the Aral Sea Region.

RESULTS

As a result, the change of farmland is as follows: the farmland area of the Aral Sea Region has abandoned 3,129 km² from 1987 to 2019, with overall accuracy of 85.3%. The farmland change had increased the drainage downstream of the Amu Darya River and the Syr Darya River. It has led area of the Aral Sea to decrease each year continuously. The area of the Aral Sea shrank by 1,606.36 km² per year from 1987 to 2019. Furthermore, Aral Sea's water level decreased by 0.13 m per year from 2003 to 2009. The amount of water storage in the Aral Sea Region also showed a downward trend from 2002 to 2016. There was a high-quality coupling coordination 0.903 relationship between surface water and farmland. It will increase the burden of water for people's normal daily life by the water loss resources caused by abandoned farmland. This study emphasized threat of unreasonable farmland management to surface water of the Aral Sea Region. The findings contributed for decision makers to formulating effective reasonable policies to protect surface water and use land of the Aral Sea Region. Meanwhile, the application of coupling degree model can provide a new method for studying the connection of independent systems in the farmland, water, environment and more.

Data-driven method based on deep learning algorithm for detecting fat, oil, and grease (FOG) of sewer networks in urban commercial areas

The content of fat, oil and grease (FOG) in the sewer network sediments is the key indicator for diagnosing sewer blockage and overflow. However, the traditional FOG detection is time-consuming and costly, and the establishment of mathematical models based on statistical methods to predict the content of FOG fail to provide satisfactory accuracy. Herein, a deep learning algorithm used a data-driven FOG content prediction model is proposed to achieve a more accurate prediction of FOG content. Meanwhile, global sensitivity analysis (GSA) is exploited to evaluate the contribution of input indicators to the output indicator (FOG) in the model, so that some input indicators that have less impact on the prediction performance can be screened out, the best combination of input indicators can be determined, and the operation cost of the model can be reduced. To evaluate the effectiveness of the proposed model, a case study was conducted in a city in southern China. The experimental results indicate that the prediction model obtains good FOG estimations and performs well from a single site to multiple sites with a mean R² of 0.922, showing a good generalization performance. Through GSA, the key input indicators in the model were identified as pH, water temperature (T), relative humidity (RH), sewage flow (Flow), drinking water supply (DWS), velocity (V) and conductivity (σ), and the input indicators such as air pressure (AP), population (Pop.), and liquid level (LV) can be reduced without affecting the prediction accuracy of the model.

Exploring spatiotemporal variation characteristics of China's industrial carbon emissions on the basis of multi-source data

Spatiotemporal variations of industrial carbon emissions (IE) must be scientifically understood, which will be helpful to formulate reasonable emission reduction strategies. Given that spatial distribution of IE is irrelevant to space agents commonly used (such as population and nighttime light), estimation and spatialization methods for total carbon dioxide (CO₂) emissions are not entirely suitable for IE. Therefore, this paper used greenhouse gases observing satellite level 4A product to estimate IE at the city level and used industrial land density to obtain the distribution of IE within the administrative districts. Sectoral emission inventories of 182 cities and a mosaic Asian anthropogenic emission inventory named MIX were used to verify the results. Then, spatiotemporal variation characteristics of China's IE were analyzed from multiple levels. Results showed that (1) the mean relative error of estimation results was 56.11%, among which 62 cities had relative error of less than 30%. Gridded IE in this paper had high consistency with MIX. (2) Cities with high IE experienced rapid growth from 2009 to 2012, followed by slower growth from 2012 to 2017. (3) Centroid of significant cold and hot spots moved to the southeast and northwest, respectively. Most cities with high annual IE growth had relatively low emission efficiency, mainly located in Inner Mongolia and Xinjiang. Aggregation of medium and high IE grids may represent high emission efficiency. Significant differences still exist between cities in IE, and sustainable development strategies should be formulated according to local conditions. Regions with high annual growth or low emission efficiency are the key to achieving IE reduction targets in future.

Multi-source data-based spatial variations of blue and green water footprints for rice production in Jilin Province, China

Rice production consumes more water than the production of other crop species due to the specific growth requirements of this species. Accurately accounting for water consumption during rice production and analyzing the spatio-temporal changes in water consumption are thus necessary. Using the water footprint (WF) as an indicator and combining data from multi-sources, this paper explored the regional differences in rice WFs in Jilin Province at a spatial resolution of 1 km. The results showed that the blue WF was always larger than the green WF, and the total, green and blue WFs were lowest during the humid year. The pixels with high values of total, green and blue WFs were mainly distributed in the eastern region of Jilin Province. Compared with the traditional estimation of the WF based on the data of administrative regions, RS techniques can overcome the administrative boundary and provide near real-time data concerning specific agricultural parameters to extract more accurate results for WF models. The combination of RS data and statistical, observational, and survey data can thus overcome the limitations of weather conditions affecting RS, reduce the incorporation of parameters, and estimate WFs quickly and accurately. This study provides a framework to evaluate crop WFs with multi-source data.

Spatiotemporal variability of glacier changes and their controlling factors in the Kanchenjunga region, Himalaya based on multi-source remote sensing data from 1975 to 2015

Spatio-temporal behavior of glaciers in the Himalayas has varied greatly in response to reported climate warming and other modulating factors such as topography, debris cover, and glacier morphology. In this paper, 429 glaciers were examined in the Kanchenjunga region in the middle of the Himalayas. Geodetic methods, feature-based image matching, and multi-parametric integrated approaches were used to detect differences of glacier change and the dominant characteristics driving these differences based on digital elevation models (DEMs), Landsat TM/ETM+/OLI images, Envisat/ASAR and Sentinel-1 data. The results showed that the average change rates in glacier area and surface elevation in 1975-2015 were -0.18 ± 0.07% a^-1 and - 0.32 ± 0.02 m a^-1, respectively. The rates of areal shrinkage of glaciers and the glacier surface velocity on the northern side of the Himalayan crest were 1.25 and 1.7 times larger than those of the glaciers on the southern slopes, respectively, whereas the rates of glacier thinning were lower in the north than in the south. The temperature increase from 1975 to 2015 caused an overall widespread glacier retreat in the region. However, differences in the topography of the Kanchenjunga region led to spatial variability in glacier changes with discrepancies as large as several times. The features of individual glaciers, such as glacier size, debris cover, and development of ice-contact glacial lakes enhanced the local complexity of glacier change and elusive response behaviors of the glaciers to climate warming led by the different topographic conditions.

The effectiveness of quarantine and isolation determine the trend of the COVID-19 epidemics in the final phase of the current outbreak in China

•

Since January 23rd 2020, stringent measures for controlling the novel coronavirus epidemics have been enforced and strengthened in mainland China.

•

Most infected cases have been quarantined or put in suspected class, which has been ignored in existing models.

•

Results of our model show that the trend of the epidemics mainly depends on quarantined and suspected cases.

•

It is important to continue enhancing the quarantine and isolation strategy and improving the detection rate in mainland China.

Objectives

Since January 23rd 2020, stringent measures for controlling the novel coronavirus epidemics have been gradually enforced and strengthened in mainland China. The detection and diagnosis have been improved as well. However, the daily reported cases staying in a high level make the epidemics trend prediction difficult.

Methods

Since the traditional SEIR model does not evaluate the effectiveness of control strategies, a novel model in line with the current epidemics process and control measures was proposed, utilizing multisource datasets including cumulative number of reported, death, quarantined and suspected cases.

Results

Results show that the trend of the epidemics mainly depends on quarantined and suspected cases. The predicted cumulative numbers of quarantined and suspected cases nearly reached static states and their inflection points have already been achieved, with the epidemics peak coming soon. The estimated effective reproduction numbers using model-free and model-based methods are decreasing, as well as new infections, while new reported cases are increasing. Most infected cases have been quarantined or put in suspected class, which has been ignored in existing models.

Conclusions

The uncertainty analyses reveal that the epidemics is still uncertain and it is important to continue enhancing the quarantine and isolation strategy and improving the detection rate in mainland China.

A spatiotemporal deep learning approach for citywide short-term crash risk prediction with multi-source data

The primary objective of this study is to investigate how the deep learning approach contributes to citywide short-term crash risk prediction by leveraging multi-source datasets. This study uses data collected from Manhattan in New York City to illustrate the procedure. The following multiple datasets are collected: crash data, large-scale taxi GPS data, road network attributes, land use features, population data and weather data. A spatiotemporal convolutional long short-term memory network (STCL-Net) is proposed for predicting the citywide short-term crash risk. A total of nine prediction tasks are conducted and compared, including weekly, daily and hourly models with 8 × 3, 15 × 5 and 30 × 10 grids, respectively. The results suggest that the prediction performance of the proposed model decreases as the spatiotemporal resolution of prediction task increases. Moreover, four commonly-used econometric models, and four state-of-the-art machine-learning models are selected as benchmark methods to compare with the proposed STCL-Net for all the crash risk prediction tasks. The comparative analyses suggest that in general the proposed STCL-Net outperforms the benchmark methods for different crash risk prediction tasks in terms of higher prediction accuracy rate and lower false alarm rate. The results verify that the proposed spatiotemporal deep learning approach performs better at capturing the spatiotemporal characteristics for the citywide short-term crash risk prediction. In addition, the comparative analyses also reveal that econometric models perform better than machine-learning models in weekly crash risk prediction tasks, while they exhibit worse results than machine-learning models in daily crash risk prediction tasks. The results can potentially guide transportation safety engineers to select appropriate methods for different crash risk prediction tasks.

A New Stochastic Kriging Method for Modeling Multi-Source Exposure-Response Data in Toxicology Studies

One of the most fundamental steps in risk assessment is to quantify the exposure-response relationship for the material/chemical of interest. This work develops a new statistical method, referred to as SKQ (stochastic kriging with qualitative factors), to synergistically model exposure-response data, which often arise from multiple sources (e.g., laboratories, animal providers, and shapes of nanomaterials) in toxicology studies. Compared to the existing methods, SKQ has several distinct features. First, SKQ integrates data across multiple sources and allows for the derivation of more accurate information from limited data. Second, SKQ is highly flexible and able to model practically any continuous response surfaces (e.g., dose-time-response surface). Third, SKQ is able to accommodate variance heterogeneity across experimental conditions and to provide valid statistical inference (i.e., quantify uncertainties of the model estimates). Through empirical studies, we have demonstrated SKQ's ability to efficiently model exposure-response surfaces by pooling information across multiple data sources. SKQ fits into the mosaic of efficient decision-making methods for assessing the risk of a tremendously large variety of nanomaterials and helps to alleviate safety concerns regarding the enormous amount of new nanomaterials.