A Machine Learning-Based Classification System for Urban Built-Up Areas Using Multiple Classifiers and Data Sources

Using POI and time series Landsat data to identify and rebuilt surface mining, vegetation disturbance and land reclamation process based on Google Earth Engine

The development of coal resources is necessary, but it has a huge negative impact on land, ecology, and the environment. With the increasing awareness of environmental protection and the requirements of related regulations, the design and practice of reclamation projects run through the mining life cycle and continue for a long time after the coal production. High-precision monitoring of mining disturbance and reclamation, quantifying the degree and time of vegetation disturbance and restoration, is of great significance to minimize the environmental effect of mining. Remote sensing, widely used as efficient monitoring tool, but there is not enough research on disturbance and reclamation monitoring taking into account large-scale areas and high temporal and spatial accuracy. Especially when mining sites remain unknown, how to distinguish the disturbance of coal mining and other human activities affecting the surface land cover has become a challenge. Therefore, this paper proposed a method to reconstruct the time series of mining disturbance and reclamation in a large area by using the POI (point of interest) and Landsat time series images using multiple buffer analysis methods. The process includes: (1) Retrieval of POI in the study area based on the public mining list using Python crawler, and buffering 100 km for preliminary extraction of potential mining areas; (2) Using spectral index mask and random forest algorithm to accurately extract the exposed coal on the Google Earth Engine (GEE) platform; (3) Buffering 10 km to identify the occurrence of disturbance and reclamation, using pixel-based temporal trajectory identification of LandTrendr algorithm under GEE. The method successful detect the change points of surface coal mining disturbance and reclamation in eastern Inner Mongolia of China. The results show that: (1) The method can effectively identify the extent of surface coal mining disturbance and reclamation, and the overall extraction accuracy is 81%. (2) Surface coal mining disturbance in eastern Inner Mongolia was concentrated in 2006-2011. By 2020, the total disturbed area is 627.8 km², with an average annual disturbance of 18.5 km², and the annual maximum disturbance to the ground reached 64.6 km² in 2008. With the total reclaimed area being 236.3 km², the reclamation rate is about 37.6%. This study provides a systematic solution and process for monitoring the disturbance and reclamation of surface coal mining in a large range with little known about the mines' location. It can effectively identify the mining disturbance and reclamation process which can also be extended to other areas, providing a quantitative assessment of mining disturbance and reclamation, which can support further ecological restoration decision-making.

Development and Validation of an Insulin Resistance Model for a Population with Chronic Kidney Disease Using a Machine Learning Approach

Background: Chronic kidney disease (CKD) is a complex syndrome without a definitive treatment. For these patients, insulin resistance (IR) is associated with worse renal and patient outcomes. Until now, no predictive model using machine learning (ML) has been reported on IR in CKD patients. Methods: The CKD population studied was based on results from the National Health and Nutrition Examination Survey (NHANES) of the USA from 1999 to 2012. The homeostasis model assessment of IR (HOMA-IR) was used to assess insulin resistance. We began the model building process via the ML algorithm (random forest (RF), eXtreme Gradient Boosting (XGboost), logistic regression algorithms, and deep neural learning (DNN)). We compared different receiver operating characteristic (ROC) curves from different algorithms. Finally, we used SHAP values (SHapley Additive exPlanations) to explain how the different ML models worked. Results: In this study population, 71,916 participants were enrolled. Finally, we analyzed 1,229 of these participants. Their data were segregated into the IR group (HOMA IR > 3, n = 572) or non-IR group (HOMR IR ≤ 3, n = 657). In the validation group, RF had a higher accuracy (0.77), specificity (0.81), PPV (0.77), and NPV (0.77). In the test group, XGboost had a higher AUC of ROC (0.78). In addition, XGBoost also had a higher accuracy (0.7) and NPV (0.71). RF had a higher accuracy (0.7), specificity (0.78), and PPV (0.7). In the RF algorithm, the body mass index had a much larger impact on IR (0.1654), followed by triglyceride (0.0117), the daily calorie intake (0.0602), blood HDL value (0.0587), and age (0.0446). As for the SHAP value, in the RF algorithm, almost all features were well separated to show a positive or negative association with IR. Conclusion: This was the first study using ML to predict IR in patients with CKD. Our results showed that the RF algorithm had the best AUC of ROC and the best SHAP value differentiation. This was also the first study that included both macronutrients and micronutrients. We concluded that ML algorithms, particularly RF, can help determine risk factors and predict IR in patients with CKD.

Analysis of the Spatio-Temporal Patterns of Shrinking Cities in China: Evidence from Nighttime Light

Since the 1980s, rapid urbanization in China has been accompanied by city shrinkage. Identifying shrinking cities and clarifying the spatial and temporal patterns are of great significance for formulating policies and realizing smart shrinkage. City shrinkage characterized by population loss is a difficult challenge for urban planning and regional development policy-making. This paper uses 2012–2020 nighttime light (NTL) data to identify the spatial and temporal distribution patterns of shrinking cities in China and excavates the shrinking cities’ trend of agglomeration and dispersion further. The following results are obtained. (1) About 34.9% of prefecture-level cities are shrinking across the country but most severely in northeast and northwest China; (2) the number of shrinking cities fluctuates over time (2015 and 2020 are the peak shrinkage years). Shrinking cities in China show a northeast-to-southwest spatial distribution. (3) From 2012 to 2020, the aggregation degree of shrinkage continuously decreased (Low-Low) and the aggregation degree of growth continuously increased (High-High), indicating that shrinkage in northeast China was slightly alleviated and that the radiative effect of the growth pole was further enhanced. These findings help us better understand the trend of city shrinkage in China. Future work needs to be focused on the potential causes of the shrinkage. Furthermore, long-term trends also need to be investigated.

Effectiveness of Semi-Supervised Learning and Multi-Source Data in Detailed Urban Landuse Mapping with a Few Labeled Samples

Detailed urban landuse information plays a fundamental role in smart city management. A sufficient sample size has been identified as a very crucial pre-request in machine learning algorithms for urban landuse classification. However, it is often difficult to recognize and label landuse categories from remote sensing images alone. Alternatively, field investigation is time-consuming with a high demand in human resources and monetary cost. Therefore, previous studies on urban landuse classification have often relied on a small size of labeled samples with very uneven spatial distribution. This study aims to explore the effectiveness of a semi-supervised classification framework with multi-source data for detailed urban landuse classification with a few labeled samples. A disagreement-based semi-supervised learning approach, the co-forest, was employed and compared with traditional supervised methods (e.g., random forest and XGBoost). Multi-source geospatial data were utilized including optical and nighttime light remote sensing and geospatial big data, which present the physical and socio-economic features of landuse categories. Taking urban landuse classification in Shenzhen City as a case, results show that the classification accuracy of the semi-supervised method are generally on par with that of traditional supervised methods, and less labeled samples are needed to achieve a comparable result under different training set ratios. Given a small sample size, the accuracy tends to be stable with training samples no less than 5% in total. Our results also indicate that the classification accuracy by using multi-source data is significantly higher than that with any single data source being applied. Among these data, map POI and high-resolution optical remote sensing data make larger contributions on the classification, followed by mobile data and nighttime light remote sensing data.

Development and Validation of an Insulin Resistance Predicting Model Using a Machine-Learning Approach in a Population-Based Cohort in Korea

Background: Insulin resistance is a common etiology of metabolic syndrome, but receiver operating characteristic (ROC) curve analysis shows a weak association in Koreans. Using a machine learning (ML) approach, we aimed to generate the best model for predicting insulin resistance in Korean adults aged > 40 of the Ansan/Ansung cohort using a machine learning (ML) approach. Methods: The demographic, anthropometric, biochemical, genetic, nutrient, and lifestyle variables of 8842 participants were included. The polygenetic risk scores (PRS) generated by a genome-wide association study were added to represent the genetic impact of insulin resistance. They were divided randomly into the training (n = 7037) and test (n = 1769) sets. Potentially important features were selected in the highest area under the curve (AUC) of the ROC curve from 99 features using seven different ML algorithms. The AUC target was ≥0.85 for the best prediction of insulin resistance with the lowest number of features. Results: The cutoff of insulin resistance defined with HOMA-IR was 2.31 using logistic regression before conducting ML. XGBoost and logistic regression algorithms generated the highest AUC (0.86) of the prediction models using 99 features, while the random forest algorithm generated a model with 0.82 AUC. These models showed high accuracy and k-fold values (>0.85). The prediction model containing 15 features had the highest AUC of the ROC curve in XGBoost and random forest algorithms. PRS was one of 15 features. The final prediction models for insulin resistance were generated with the same nine features in the XGBoost (AUC = 0.86), random forest (AUC = 0.84), and artificial neural network (AUC = 0.86) algorithms. The model included the fasting serum glucose, ALT, total bilirubin, HDL concentrations, waist circumference, body fat, pulse, season to enroll in the study, and gender. Conclusion: The liver function, regular pulse checking, and seasonal variation in addition to metabolic syndrome components should be considered to predict insulin resistance in Koreans aged over 40 years.

Evaluation of Polycentric Spatial Structure in the Urban Agglomeration of the Pearl River Delta (PRD) Based on Multi-Source Big Data Fusion

The rapid development of the urban city has led to great changes in the urban spatial structure. Thus, analyses of polycentric urban spatial structures are important for understanding these kinds of structures. In order to accurately evaluate the polycentric spatial structure of urban agglomerations and judge the differences between the actual development situation and overall planning of urban agglomerations, this study proposes a new method to identify the polycentric spatial structure of urban agglomerations in the Pearl River Delta based on the fusion of nighttime light (NTL) data, point of interest (POI) data, and Tencent migration data (TMG). In the first step, the NTL, POI, and TMG data are fused via wavelet transform; in the second step, Anselin local Moran’s I (LMI) and geographically weighted regression (GWR) were used to identify the main centers and subcenters, respectively. In the third step, the accuracy of the results of this study was further verified and discussed in the context of overall planning. The results show that the accuracy of urban polycenter identification via LMI and GWR after data fusion was 92.84%, and the Kappa value was 0.8971, which was higher than the results of polycenter identification via the traditional relative threshold. After comparing the identification results with the overall planning, firstly, we see that the fusion of multi-source big data can help to accurately evaluate the polycentric spatial structure within the urban agglomeration. Secondly, the fusion of dynamic data and static data can help identify the polycentric spatial structure of urban space more accurately. Therefore, this study can provide a new design for urban polycentric spatial structures, and further provide a reliable reference for the spatial optimization of urban agglomeration and the formulation of regional spatial development policies.

Extraction of urban built-up area based on the fusion of night-time light data and point of interest data

The accurate extraction of urban built-up areas is an important prerequisite for urban planning and construction. As a kind of data that can represent urban spatial form, night-time light data has been widely used in the extraction of urban built-up areas. As one of the geographic open-source big data, point of interest (POI) data has a high spatial coupling with night-time light data, so researchers are beginning to explore the fusion of the two data in order to achieve more accurate extraction of urban built-up areas. However, the current research methods and theoretical applications of the fusion of POI data and night-time light data are still insufficient compared with the dramatically changing urban built-up areas, which needed to be further supplemented and deepened. This study proposes a new method to fuse POI data and night-time light data. The results before and after data fusion are compared, and the accuracy of urban built-up area extracted by different data and methods is analysed. The results show that the data fusion can avoid the shortage of single data and effectively improve the extraction accuracy of urban built-up areas, which is greatly helpful to supplement the study of data fusion in urban built-up areas, and also can provide decision-making guidance for urban planning and construction.

COVID-19 Risk Assessment: Contributing to Maintaining Urban Public Health Security and Achieving Sustainable Urban Development

As the most infectious disease in 2020, COVID-19 is an enormous shock to urban public health security and to urban sustainable development. Although the epidemic in China has been brought into control at present, the prevention and control of it is still the top priority of maintaining public health security. Therefore, the accurate assessment of epidemic risk is of great importance to the prevention and control even to overcoming of COVID-19. Using the fused data obtained from fusing multi-source big data such as POI (Point of Interest) data and Tencent-Yichuxing data, this study assesses and analyzes the epidemic risk and main factors that affect the distribution of COVID-19 on the basis of combining with logistic regression model and geodetector model. What’s more, the following main conclusions are obtained: the high-risk areas of the epidemic are mainly concentrated in the areas with relatively dense permanent population and floating population, which means that the permanent population and floating population are the main factors affecting the risk level of the epidemic. In other words, the reasonable control of population density is greatly conducive to reducing the risk level of the epidemic. Therefore, the control of regional population density remains the key to epidemic prevention and control, and home isolation is also the best means of prevention and control. The precise assessment and analysis of the epidemic conducts by this study is of great significance to maintain urban public health security and achieve the sustainable urban development.

Quantifying Spatiotemporal Changes in Human Activities Induced by COVID-19 Pandemic Using Daily Nighttime Light Data

The COVID-19 pandemic caused drastic changes in human activities and nighttime light (NTL) at various scales, providing a unique opportunity for exploring the pattern of the extreme responses of human community. This study used daily NTL data to examine the spatial variations and temporal dynamics of human activities under the influence of COVID-19, taking Chinese mainland as the study area. The results suggest that the change in the intensity of NTL is not correlated to the number of confirmed cases, but reflects the changes in human activities and the intensity of epidemic prevention and control measures within a region. During the outbreak period, the major provincial capitals and urban agglomerations were affected by COVID-19 more than smaller cities. During the recovery, different regions showed different recovery processes. The cities in West and Northeast China recovered steadily while the recovery in coastal cities showed relatively greater fluctuations due to an increase in imported cases. Wuhan, the most seriously affected city in China, did not recover until the end of March. Nevertheless, as of 31 March, the overall NTL across China had recovered to an 89.5% level of the same period in the previous year. The high consistency between the big data of travel intensity and NTL further proved the validity of the results of this study. These findings imply that daily NTL data are effective for rapidly monitoring the dynamic changes in human activities, and can help evaluate the effects of control measures on human activities during major public health events.

Using Wavelet Transforms to Fuse Nighttime Light Data and POI Big Data to Extract Urban Built-Up Areas

Urban built-up areas are not only the embodiment of urban expansion but also the main space carrier of urban activities. Accurate extraction of urban built-up areas is of great practical significance for measuring the urbanization process and judging the urban environment. It is difficult to identify urban built-up areas objectively and accurately with single data. Therefore, to evaluate urban built-up areas more accurately, this study uses the new method of fusing wavelet transforms and images on the basis of utilization of the POI data of March 2019 and the Luojia1-A data from October 2018 to March 2019. to identify urban built-up areas. The identified urban built-up areas are mainly concentrated in the areas with higher urbanization level and night light value, such as the northeast of Dianchi Lake and the eastern bank around the Dianchi Lake. It is shown in the accuracy verification result that the classification accuracy identified by night-light data of urban build-up area accounts for 84.00% of the total area with the F1 score 0.5487 and the Classification accuracy identified by the fusion of night-light data and POI data of urban build-up area accounts for 96.27% of the total area with the F1 score 0.8343. It is indicated that the built-up areas identified after image fusion are significantly improved with more realistic extraction results. In addition, point of interest (POI) data can better account for the deficiency in nighttime light (NTL) data extraction of urban built-up areas in the urban spatial structure, making the extraction results more objective and accurate. The method proposed in this study can extract urban built-up areas more conveniently and accurately, which is of great practical significance for urbanization monitoring and sustainable urban planning and construction.