Highest Kappa Coefficient Research Articles

Urban Catchment-Scale Blue-Green-Gray Infrastructure Classification with Unmanned Aerial Vehicle Images and Machine Learning Algorithms

Green infrastructure (GI), such as green roofs, is now widely used in sustainable urban development. An accurate mapping of GI is important to provide surface parameterization for model development. However, the accuracy and precision of mapping GI is still a challenge in identifying GI at the small catchment scale. We proposed a framework for blue-green-gray infrastructure classification using machine learning algorithms and unmanned aerial vehicle (UAV) images that contained digital surface model (DSM) information. We used the campus of the Southern University of Science and Technology in Shenzhen, China, as a study case for our classification method. The UAV was a DJI Phantom 4 Multispectral, which measures the blue, green, red, red-edge, and near-infrared bands and DSM information. Six machine learning algorithms, i.e., fuzzy classifier, k-nearest neighbor classifier, Bayes classifier, classification and regression tree, support vector machine (SVM), and random forest (RF), were used to classify blue (including water), green (including green roofs, grass, trees (shrubs), bare land), and gray (including buildings, roads) infrastructure. The highest kappa coefficient was observed for RF and the lowest was observed for SVM, with coefficients of 0.807 and 0.381, respectively. We optimized the sampling method based on a chessboard grid and got the optimal sampling interval of 11.6 m to increase the classification efficiency. We also analyzed the effects of weather conditions, seasons, and different image layers, and found that images in overcast days or winter days could improve the classification accuracy. In particular, the DSM layer was crucial for distinguishing green roofs and grass, and buildings and roads. Our study demonstrates the feasibility of using UAV images in urban blue-green-gray infrastructure classification, and our infrastructure classification framework based on machine learning algorithms is effective. Our results could provide the basis for the future urban stormwater management model development and aid sustainable urban planning.

Recognition of Areca Leaf Yellow Disease Based on PlanetScope Satellite Imagery

Areca yellow leaf disease is a major attacker of the planting and production of arecanut. The continuous expansion of arecanut (Areca catechu L.) planting areas in Hainan has placed a great need to strengthen the monitoring of this disease. At present, there is little research on the monitoring of areca yellow leaf disease. PlanetScope imagery can achieve daily global coverage at a high spatial resolution (3 m) and is thus suitable for the high-precision monitoring of plant pest and disease. In this paper, PlanetScope images were employed to extract spectral features commonly used in disease, pest and vegetation growth monitoring for primary models. In this paper, 13 spectral features commonly used in vegetation growth and pest monitoring were selected to form the initial feature space, followed by the implementation of the Correlation Analysis (CA) and independent t-testing to optimize the feature space. Then, the Random Forest (RF), Backward Propagation Neural Network (BPNN) and AdaBoost algorithms based on feature space optimization to construct double-classification (healthy, diseased) monitoring models for the areca yellow leaf disease. The results indicated that the green, blue and red bands, and plant senescence reflectance index (PSRI) and enhanced vegetation index (EVI) exhibited highly significant differences and strong correlations with healthy and diseased samples. The RF model exhibits the highest overall recognition accuracy for areca yellow leaf disease (88.24%), 2.95% and 20.59% higher than the BPNN and AdaBoost models, respectively. The commission and omission errors were lowest with the RF model for both healthy and diseased samples. This model also exhibited the highest Kappa coefficient at 0.765. Our results exhibit the feasible application of PlanetScope imagery for the regional large-scale monitoring of areca yellow leaf disease, with the RF method identified as the most suitable for this task. Our study provides a reference for the monitoring, a rapid assessment of the area affected and the management planning of the disease in the agricultural and forestry industries.

A Prediction Method of Electromagnetic Environment Effects for UAV LiDAR Detection System

With the rapid development of science and technology, UAVs (Unmanned Aerial Vehicles) have become a new type of weapon in the informatization battlefield by their advantages of low loss and zero casualty rate. In recent years, UAV navigation electromagnetic decoy and electromagnetic interference crashes have activated widespread international attention. The UAV LiDAR detection system is susceptible to electromagnetic interference in a complex electromagnetic environment, which results in inaccurate detection and causes the mission to fail. Therefore, it is very necessary to predict the effects of the electromagnetic environment. Traditional electromagnetic environment effect prediction methods mostly use a single model of mathematical model and machine learning, but the traditional prediction method has poor processing nonlinear ability and weak generalization ability. Therefore, this paper uses the Stacking fusion model algorithm in machine learning to study the electromagnetic environment effect prediction. This paper proposes a Stacking fusion model based on machine learning to predict electromagnetic environment effects. The method consists of Extreme Gradient Boosting algorithm (XGB), Gradient Boosting Decision Tree algorithm (GBDT), K Nearest Neighbor algorithm (KNN), and Decision Tree algorithm (DT). Experimental results show that, comprising with the other seven machine learning algorithms, the Stacking fusion model has a better classification prediction accuracy of 0.9762, a lower Hamming code distance of 0.0336, and a higher Kappa coefficient of 0.955. The fusion model proposed in this paper has a better predictive effect on electromagnetic environment effects and is of great significance for improving the accuracy and safety of UAV LiDAR detection systems under the complex electromagnetic environment on the battlefield.

Identification of cultivated land change trajectory and analysis of its process characteristics using time-series Landsat images: A study in the overlapping areas of crop and mineral production in Yanzhou City, China

Most studies have extensively evaluated the extent and direction of land-use change in coal mining areas; however, they did not adequately describe the time dimension of cultivated land changes at the pixel scale. In this study, we reconstructed the time-series of the normalized difference vegetation index (NDVI) using best index slope extraction-wavelet transform (BISE-WT) filtering. The trajectory type of cultivated land change was identified based on the time-series curves of those original cultivated land pixels using the modified normalized difference water index (MNDWI), normalized differences building index (NDBI), and bare soil index (BSI). Additionally, the time nodes of cultivated land changes were detected based on the NDVI time-series data, MNDWI, NDBI, and BSI. The results showed that this clustering method had the highest overall accuracy (89.90%) and the highest kappa coefficient (86.36%) of those three methods. Moreover, the overall accuracy of different trajectory types, time node detection in cultivated land converted to other lands, and the restored cultivated land from other lands were 0.9005, 0.9438, and 0.9430, respectively, and the kappa coefficient were 0.8803, 0.9390, and 0.9371, respectively. The conversion from cultivated land to non-cultivated land mainly occurred during 1989–2005, while the reclamation of cultivated land mainly occurred in 2009, 2011, and 2013. Permanent cultivated land accounted for the highest proportion (56.26%) of the five trajectory types. The proportion of cultivated land converted to non-cultivated land to cultivated land was 18.51%, and the proportion of disturbed cultivated land that was not reclaimed was 25.23%. The proportion of cultivated land converted to the developed was the comparatively high (17.73%), and that of the restored cultivated land after conversion from cultivated land to waterbody was the lowest (0.53%). The results of this study provide a scientific basis for guiding land reclamation, ecological restoration, and evaluating sustainability in the overlapping areas of crop and mineral production.

Land cover classification in a mixed forest-grassland ecosystem using LResU-net and UAV imagery

Using an unmanned aerial vehicle (UAV) paired with image semantic segmentation to classify land cover within natural vegetation can promote the development of forest and grassland field. Semantic segmentation normally excels in medical and building classification, but its usefulness in mixed forest-grassland ecosystems in semi-arid to semi-humid climates is unknown. This study proposes a new semantic segmentation network of LResU-net in which residual convolution unit (RCU) and loop convolution unit (LCU) are added to the U-net framework to classify images of different land covers generated by UAV high resolution. The selected model enhanced classification accuracy by increasing gradient mapping via RCU and modifying the size of convolution layers via LCU as well as reducing convolution kernels. To achieve this objective, a group of orthophotos were taken at an altitude of 260 m for testing in a natural forest-grassland ecosystem of Keyouqianqi, Inner Mongolia, China, and compared the results with those of three other network models (U-net, ResU-net and LU-net). The results show that both the highest kappa coefficient (0.86) and the highest overall accuracy (93.7%) resulted from LResU-net, and the value of most land covers provided by the producer’s and user’s accuracy generated in LResU-net exceeded 0.85. The pixel-area ratio approach was used to calculate the real areas of 10 different land covers where grasslands were 67.3%. The analysis of the effect of RCU and LCU on the model training performance indicates that the time of each epoch was shortened from U-net (358 s) to LResU-net (282 s). In addition, in order to classify areas that are not distinguishable, unclassified areas were defined and their impact on classification. LResU-net generated significantly more accurate results than the other three models and was regarded as the most appropriate approach to classify land cover in mixed forest-grassland ecosystems.

DNA microarray analysis of Staphylococcus aureus from Nigeria and South Africa.

Staphylococcus aureus is an important human pathogen with an arsenal of virulence factors and a propensity to acquire antibiotic resistance genes. The understanding of the global epidemiology of S. aureus through the use of various typing methods is important in the detection and tracking of novel and epidemic clones in countries and regions. However, detailed information on antibiotic resistance and virulence genes of S. aureus, and its population structure is still limited in Africa. In this study, S. aureus isolates collected in South Africa (n = 38) and Nigeria (n = 2) from 2001-2004 were characterized by spa typing and DNA microarray. The combination of these two methods classified the isolates into seven spa types and three clonal complexes (CCs) i.e. t064-CC8 (n = 17), t037-CC8 (n = 8), t1257-CC8 (n = 6), t045-CC5 (n = 5), t951-CC8 (n = 1), t2723-CC88 (n = 1), t6238-CC8 (n = 1), and untypeable-CC8 (n = 1). A high percentage agreement (>95%) and kappa coefficient (>0.60) was largely observed with antibiotic susceptibility testing and DNA microarray, indicating substantial agreement. Some antibiotic and virulence gene markers were associated with specific clones. The detection of the collagen-binding adhesion (cna) gene was unique for t037-CC8-MRSA while the enterotoxin gene cluster (egc) and staphylococcal complement inhibitor (scn) gene were identified with t045-CC5-MRSA. Moreover, the combination of genes encoding enterotoxins (entA, entB, entK, entQ) was noted with most of the CC8 isolates. The t045-CC5-MRSA clone was positive for the mercury resistance (mer) operon. DNA microarray provides information on antibiotic resistance and virulence gene determinants and can be a useful tool to identify gene markers for specific S. aureus clones in Africa.

Impact Assessments of Typhoon Lekima on Forest Damages in Subtropical China Using Machine Learning Methods and Landsat 8 OLI Imagery

Wind damage is one of the major factors affecting forest ecosystem sustainability, especially in the coastal region. Typhoon Lekima is among the top five most devastating typhoons in China and caused economic losses totaling over USD 8 billion in Zhejiang Province alone during 9–12 August 2019. However, there still is no assessment of its impacts on forests. Here we detected forest damage and its spatial distribution caused by Typhoon Lekima by classifying Landsat 8 OLI images using the random forest (RF) machine learning algorithm and the univariate image differencing (UID) method on the Google Earth Engine (GEE) platform. The accuracy assessment indicated a high overall accuracy (>87%) and kappa coefficient (>0.75) for forest-damage detection, as evaluated against field-investigated plot data, with better performance using the RF method. The total affected forest area by Lekima was 4598.87 km2, accounting for 8.44% of the total forest area in Zhejiang Province. The light-, moderate- and severe-damage forest areas were 2106.29 km2, 2024.26 km2 and 469.76 km2, respectively. Considering the damage severity, the net forest canopy loss fraction was 2.57%. The affected forest area and damage severity exhibited large spatial variations, which were affected by elevation, slope, precipitation and forest type. Our study indicated a larger uncertainty for affected forest area and a smaller uncertainty for the proportion of damage severity, based on multiple assessment approaches. This is among the first studies on forest damage due to typhoons at a regional scale in China, and the methods can be extended to examine the impacts of other super-strong typhoons on forests. Our study results on damage severity, spatial distribution and controlling factors could help local governments, the forest sector and forest landowners make decision on tree-planting planning and sustainable management after typhoon strikes and could also raise public and governmental awareness of typhoons’ damage on China’s inland forests.

Alpha-defensins determination in different types of synovial fluid and parallel collected serum samples by ELISA

The study aims were to verify the serum (S) and synovial fluid (SF) reference intervals (RIs) for human neutrophil defensins (HNP1-3); measure S and SF defensin concentrations in different types of SF, including non-inflammatory, inflammatory non-pyogenic, inflammatory pyogenic, and hemorrhagic; and to compare the HNP1-3 concentrations in SF and S with those of other inflammatory biomarkers. SF and S samples were collected from 92 patients. HNP1-3 concentrations were determined using enzyme-linked immunosorbent assays; glucose, lactate, interleukin-6, and procalcitonin using an automatic analyzer; and presepsin using a Pathfast system. There were 61 non-inflammatory, 11 inflammatory non-pyogenic, 11 inflammatory pyogenic, and 9 hemorrhagic SF. Non-inflammatory SF was divided into non-inflammatory normal and non-inflammatory osteoarthritis. The former was used to estimate the HNP1-3 RI in SF and S. The estimated HNP1-3 RIs of SF and S were 12.47-437.42 mg/L and 5.45-44.75 µg/L, respectively. HNP1-3 differed significantly between S and SF and individual groups of SF (P<0.001 and P=0.001, respectively). There were significant relationships between SF HNP1-3 and S HNP1-3 (P<0.001), S C-reactive protein (P<0.001), and S interleukin-6 (P=0.007), and between SF HNP1-3 and SF C-reactive protein (P=0.004) and SF interleukin-6 (P<0.001). The highest kappa coefficient was between SF HNP1-3 and SF interleukin-6 (κ=0.507). We validated the SF HNP1-3 diagnostic kit and demonstrated that SF and S HNP1-3 are promising biomarkers for distinguishing inflammatory from non-inflammatory joint diseases.

Comparison of Clinical Tests for Peripheral Diabetic Neuropathy in a Type 1 Diabetes Cohort

ObjectiveTo examine the performance and agreement of 5 modalities for testing sensory neuropathy against a neurothesiometer among Hispanic patients with type 1 diabetes (T1D) in an outpatient setting. MethodsA cross-sectional study was conducted at a tertiary reference center in Mexico City. Sensitivity, specificity, predictive values, and likelihood ratios were calculated using a VibraTip device, 128 Hz tuning fork, and the Semmes-Weinstein 5.07/10 g monofilament test, Ipswich touch test (IpTT), and pinprick test (PPT). The VPT obtained using a neurothesiometer was used as the standard. Agreement between tests was calculated using kappa coefficients. ResultsOur study included 78 patients (156 examinations), of whom 56.4% were females. The mean age was 38.2 ± 13.0 years, and the mean body mass index was 24.6 ± 4.8 kg/m2. The best sensitivity was found for IpTT and VibraTip (89.7% and 79.3%, respectively), while the PPT and IpTT had the highest positive predictive values (94.4% and 92.9%, respectively). The highest kappa coefficients were obtained for the IpTT vs neurothesiometer (kappa coefficient [κ] = 0.893, P < .001), followed by VibraTip vs neurothesiometer (κ = 0.782, P < .001). The VibraTip vs IpTT also had a substantial agreement (κ= 0.713, P < .001). ConclusionOur findings demonstrated that the IpTT had the best diagnostic performance and agreement compared with the standard in this cohort of Hispanic patients with T1D. The IpTT is a useful, simple test for diabetic neuropathy screening. These findings support its inclusion in future guidelines for diabetic foot examination.

Single-shot multibox detector- and building information modeling-based quality inspection model for construction projects

Identifying, locating, and evaluating concrete construction defects in the construction sector consumes time and is overlaid by subjectivity. This study proposes a single-shot multibox detector (SSD)-based and building information modeling (BIM)-based quality inspection framework to improve work efficiency and decrease unnecessary deviations resulting from human judgment. The SSD-based defect recognition model, which can identify surface defects with 86.56% average recall and 75.34% average precision, is trained using 1387 defect images of crack, honeycombing, water infiltration, rebar exposure, and edge or corner damage. Furthermore, a division algorithm is created to divide the BIM model surfaces into pieces, in which each piece is assigned with a unique location code number. After photographing a construction project according to the coding sequence, the defects in the images are identified and located on the BIM model. Then, the frequency of the quality defects, defect aggregation zones, and quality status in the BIM model are provided to the managers for fast-paced analysis. The case study results indicate that our inspection framework has the shortest inspection time and the highest kappa coefficients (0.816, 0.901, and 0.799) for crack, honeycombing, and reinforcement exposure prediction. The inspection accuracy of our model exceeds 98.5% for each defect. Our novel quality inspection framework can remarkably improve the efficiency and effectiveness of quality management and lead to a higher level of project management.

COMPARISON OF MACHINE LEARNING METHODS IN CLASSIFYING POVERTY IN INDONESIA IN 2018

Poverty is still one of the main problems in economic development besides inequality, unemployment, and economic growth. This study aims to model poverty directly using a discrete choice model, namely the machine learning classification method. The data used are imbalanced data where one of the categories is small enough so that the resample of both sampling method is used. In this study, several machine learning methods were applied, including the Decision Tree, Naïve Bayes, K-Nearest Neighbor (KNN), and Rotation Forest. The results show that the technique of using resample both samplings provides optimal results for the four machine learning methods. If viewed from the indicators of accuracy, specificity, sensitivity, AUC, and the highest Kappa coefficient produced, the best method is the KNN method. The KNN model has an accuracy value of 0.73 percent, sensitivity of 0.68 percent, specificity of 78 percent, and AUC of 0.73.

A Study on RB‐XGBoost Algorithm‐Based e‐Commerce Credit Risk Assessment Model

The current method’s e‐commerce credit risk assessment is prone to poor data balance and low evaluation accuracy. An RB‐XGBoost algorithm‐based e‐commerce credit risk assessment model is proposed in this study. The adaptive random balance (RB) method is used to sample and process the obtained data to improve the balance degree of the data. An assessment index system is constructed based on the processed data. Based on the risk evaluation index system and the XGBoost algorithm, this paper constructed an e‐commerce risk assessment model and assessed the e‐commerce credit risk using this model. The experimental results show that the proposed method has good data balance, a high kappa coefficient, and a large receiver operating characteristic (ROC) curve area, which can effectively improve e‐commerce credit risk assessment accuracy.

Improved Mapping of Long-Term Forest Disturbance and Recovery Dynamics in the Subtropical China Using All Available Landsat Time-Series Imagery on Google Earth Engine Platform

After implementations of many economic and forestry policies during recent 30 years, forest cover in China has experienced significant changes. The fine scale and long-term information on forest temporary (disturbance) loss, persistent loss (deforestation), and gains (afforestation/reforestation) will provide a guide for forest management and policies at regional scale. Using Jiangxi Province as a pilot study region, here we assessed forest cover dynamics in the subtropical China during 1986–2019 by integrating time-series Landsat images, Landsat-based Detection of Trends in Disturbance and Recovery algorithm and Random Forest classifier on the Google Earth Engine Platform. The accuracy assessment indicated a high overall accuracy (>90%) and Kappa coefficient (>0.9) for both forest gain and loss detections as evaluated against various sources of plot-level sample data and the existing global forest change product. The total forest loss area was 18 697.79 km2, with persistent loss area of 3394.31 km2 due to land conversion during 1986–2019, while persistent (net) forest gain area was 45 656.96 km2, accounting for 57.70% of the forest area in 1986. Forest loss area exhibited large interannual variations, but showed a general increase trend from 1986 to 2019. The annual variation patterns of forest gain and loss area were associated with the changes in forestry policies and large disturbance events. Our assessments on the long-term and fine scale forest dynamic patterns will help evaluate the effectiveness of forest management practices and forestry polices on forest resource sustainability, and climate change and greenhouse gases mitigation in Jiangxi Province and China.

Evidence-based diagnostic algorithm for visceral leishmaniasis in Bangladesh

Evidence-based diagnostic algorithm is highly recommended for the visceral leishmaniasis (VL). This cross-sectional study was performed in Bangladesh to evaluate VL diagnostic tools including serology, buffy coat smear microscopy for LD body and various DNA-based techniques using buffy coat in 100 confirmed VL cases and 100 controls. The performance of tools against spleen smear (gold standard) was evaluated using kappa coefficient. Diagnostic precision and other inherent indicators were considered for index scoring (IS) of performance of tools using factor analysis. A diagnostic algorithm was formulated based on the IS and availability of the tools at different health care facilities of Bangladesh. A high level of agreement (kappa≥ 0.80) was observed for all the diagnostic tools. The highest kappa coefficients were found for rK39 RDT and rK39 ELISA (0.95), followed by ssuRNA-PCR (0.94), Buffy coat smear (0.93), rK28 ELISA (0.92), rK28 RDT (0.89), LAMP (0.89), Mini-exon PCR (0.86), ITS1 (0.85), and ITS2 PCR (0.80). rK39 RDT was found to be the best diagnostic test (IS: 1.7) followed by rK28 RDT (IS: 1.5), buffy coat smear microscopy (IS: 0.5), rK39 & rK28 ELISA (IS: 0.3), ssuRNA-PCR (IS: -0.7) and LAMP, Mini-exon, ITS1, & ITS2 PCR (IS: -0.9). rK39 RDT has been proposed as the best option for primary health care facilities, while buffy coat smear microscopy was found to be a good adjunct for confirmation of serology-positive cases and proposed for secondary and tertiary facilities. ssuRNA-PCR or LAMP can be an alternate confirmation tool only applicable to the tertiary facilities.

Psychometric Characteristics of a Commuting-to-School Behaviour Questionnaire for Families.

The purposes of this study were: (a) to describe the patterns of modes of commuting to school (children) and to work (parents) separated by gender and age, (b) to validate the questions on children’s mode of commuting to and from school according to their parents, and (c) to analyse the reliability of a family questionnaire focused on commuting to school behaviours. A total of 611 parents (mean age: 43.28 ± 6.25 years old) from Granada (Spain) completed “Family commuting-to-school behaviour” questionnaire in two sessions separated by 14 days, (2016 and 2018). The validation between family and children’s questions was assessed using the Kappa and Spearman correlation coefficients, and the test–retest reliability within the family questions was assessed using the Kappa and the weighted Kappa. The children’s modes of commuting to school (mean age: 11.44 ± 2.77 years old) were mainly passive (57.7% to school) while parents’ modes of commuting to work were mainly active (71.6%). The validity of the mode of commuting was significant with high Kappa and Spearman coefficients. The test–retest reliability presented a good agreement for the mode of commuting to school in children, distance and time to school, and the mode of commuting to work in parents, while the questions on acceptable distance to walk or cycle to school showed a moderate to good agreement. The “Family commuting-to-school behaviour” questionnaire could be a useful tool to assess the mode of commuting of children, distance and time to school for researchers and practitioners.

Interobserver reliability of the Tile classification system for pelvic fractures among radiologists and surgeons

ObjectivesTo assess the interobserver reliability (IOR) of the Tile classification system, and its potential influence on outcomes, for the interpretation of CT images of pelvic fractures by radiologists and surgeons.MethodsRetrospective data (1/2008–12/2016) from 238 patients with pelvic fractures were analyzed. Mean patient age was 44 years (SD 20); 66% were male. There were 54 Tile A, 82 Tile B, and 102 Tile C type injuries. The 30-day mortality rate was 15% (36/238). Six observers, three radiologists, and three surgeons with different levels of experience (attending/resident/intern) classified each fracture into one of the 26 second-order subcategories of the Tile classification. Weighted kappa coefficients were used to assess the IORs for the three main categories and nine first-order subcategories.ResultsThe overall IORs of the Tile system for the main categories and first-order subcategories were moderate (kappa = 0.44) and fair (kappa = 0.31), respectively. IOR was fair to moderate among radiologists, but only fair among surgeons. By level of training, IOR was moderate between attendings and between residents, whereas it was only fair between interns. IOR was moderate to substantial (kappa = 0.56–0.70) between the radiology attending and resident. Association of the Tile fracture type with 30-day mortality was present based on two out of six observer ratings.ConclusionsThe overall IOR of the Tile classification system is only fair to moderate, increases with the level of rater experience and is better among radiologists than surgeons. In the light of these findings, results from studies using this classification system must be interpreted cautiously.Key Points• The overall interobserver reliability of the Tile pelvic fracture classification is only fair to moderate.• Interobserver reliability increases with observer experience and radiologists have higher kappa coefficients than surgeons.• Interobserver reliability has an impact on the association of the Tile classification system with mortality in two out of six cases.

High-resolution multispectral mapping facies on glacier surface in the Arctic using WorldView-3 data

Glaciers are important and sensitive part of our environment which can be used as indicator of global warming and climate change. Glacier facies represent distinct regions of a glacier surface characterized by near surface structure and density that develop as a function of spatial variations in surface melt and accumulation. The facies mapping aids in delineating different zones of the glacier, which are useful in computing glacier mass balance and modeling. In this study we tested traditional and advanced classification techniques on the Edithbreen glacier situated in Ny-lesund, Svalbard, using WorldView-3 and Landsat 8 OLI. The comparison of the accuracy was conducted using error matrices. Six measures of accuracy were derived from the error matrices and were compared with each other to find the method delivering the most adequate output for facies mapping. The pixel-based approach applied to Landsat-8 data yielded higher accuracies (&gt;80%) when compared to that. The object-oriented classification revealed a much better accuracy and high kappa coefficient for both low and high-resolution datasets. The study clearly indicates that the object-oriented classification provides better results for glacier facies classification when high spatial resolution is used, but for lower spatial resolution, pixel-based methods are adequate.

P087: A comparison of how emergency physicians and plastic surgeons evaluate and triage pediatric hand fractures: a prospective trial of the Calgary Kids’ Hand Rule

Introduction: Hand fractures in children are common and most are adequately managed with immobilization alone. There is a subset of fractures that require surgery as well as a fear of growth plate disturbance. For these reasons, triaging the so-called “complex” fractures that require specialized care by a hand surgeon is critical. In an effort to improve triaging for pediatric hand fractures, we previously derived and internally validated a prediction model for pediatric hand fracture triage using multivariable logistic regression with bootstrapping. The primary outcome was “complex fracture”, a definition we assigned to any fractures that required surgery, closed reduction or more than four appointments with a plastic surgeon. The model identified six significant predictors of complex fractures: angulation, condylar involvement, dislocation or subluxation, displacement, open fracture, and malrotation with strong performance (C-statistic 0.88) and was named the Calgary Kids’ Hand Rule (CKHR). Methods: A prospective cohort study was conducted at the Alberta Children's Hospital from April 1 until December 31, 2019. Eligible patients included children 17 years and younger with a radiographically confirmed hand fracture. Both emergency physicians and plastic surgeons completed independent CKHR forms for each new hand fracture. The fracture was predicted as “complex” if any one of the six predictors were present on the form. If none of the six predictors were present, the predicted outcome was “simple”. The observed outcome was “complex” if the fracture required surgery, closed reduction, or four or more appointments with a plastic surgeon at three months follow-up. All other fractures were observed outcome “simple”. The classification performance of the CKHR was assessed via sensitivity, specificity, and C-statistic. The kappa coefficient for inter-rater reliability between emergency physicians and plastic surgeons was calculated for each predictor. Results: To date, 102 pediatric hand fractures have been included in this prospective cohort study. Of the 74 observed simple fractures, 49 were predicted as “simple” and 25 were predicted as “complex”. Of the 28 observed “complex” fractures, 25 were predicted as “complex” and 3 were predicted as “simple”. These findings correspond to a sensitivity of 89%, specificity of 66%, and a C-statistic of 0.78. Of the 3 observed “complex” fractures that were predicted as “simple”, i.e. the 3 false negatives, one was a Seymour fracture of the fourth distal phalanx for which the emergency physician did not tick any boxes on the form. Upon further investigation, we learned that the physician had commented that the fracture was open, thus alluding to their acknowledgement of the predictor “open fracture” as being present). This fracture went on to require surgery. The second false-negative fracture was a non-displaced, intra-articular fracture of the fourth metacarpal head that required multiple appointments with the plastic surgeon. The third false-negative fracture was a non-displaced Salter-Harris II fracture of the thumb proximal phalanx without malrotation that received a closed reduction by the emergency physician. The kappa coefficient for inter-rater reliability between emergency physician and plastic surgeon evaluation of predictors varied by predictor from fair to almost perfect agreement. Condylar involvement had the highest kappa coefficient (kappa = 0.85) followed by malrotation (kappa = 0.65) and dislocation (kappa = 0.64). The predictors with the lowest kappa coefficients were displacement (kappa = 0.42), open fractures (kappa = 0.50), and angulation (kappa = 0.53). Conclusion: The Calgary Kids’ Hand Rule had a sensitivity of 89% in a prospective cohort of pediatric hand fractures referred to the Alberta Children's Hospital. The sensitivity could likely be improved with knowledge translation and specific education regarding use of the prediction tool. Emergency physicians and plastic surgeons displayed the lowest inter-rater reliability when assessing displacement, open fractures, and angulation. These predictors may represent areas of future research and physician education to delineate and decrease the discordance between emergency physicians and plastic surgeons.

Assessing 40 years of spatial dynamics and patterns in megacities along the Belt and Road region using satellite imagery

ABSTRACT The Belt and Road (B&R) region, a vital area with historical, economic, cultural and political significance, has undergone rapid urbanization in the past several decades, especially in the form of urban expansion. In this study, 20 megacities in the B&R region were selected to explore different spatiotemporal patterns of urban expansion. Object-oriented support vector machines (SVM), annual growth rate (AGR) models, and landscape metrics were employed to delineate the urban areas and characterize spatiotemporal characteristics and landscape patterns of these megacities during 1975–2015. All urban maps presented high overall accuracies (80.70%–95.90%) and overall Kappa coefficients (0.76–0.95). The study revealed that megacities in the B&R region have undergone different types of urban sprawl, mainly adopting a ‘concentric circle’ pattern in inland areas and a ‘sector’ pattern in coastal areas. Besides, six expansion modes were summarized according to the AGRs of individual megacities. Differences existed in megacities of the developing and developed countries and among five sub-regions. Moreover, ‘dispersion, gathering, and re-dispersion’ and ‘coalescence’ were two major landscape patterns of megacities in developing and developed countries. Results of this study can provide a scientific reference for urban planning and aid in sustainable development of local areas.

Appraisal of kappa-based metrics and disagreement indices of accuracy assessment for parametric and nonparametric techniques used in LULC classification and change detection

In the face of rapid urbanization, monitoring urban expansion has gained importance to sustainably manage the land resources and minimize the impact on the environment. Monitoring urban growth using satellite data involves computing the state of land use–land cover and their change over time. A number of computing methods have been developed to process and interpret the satellite results for an urban environment. However, due to a large number of parametric and nonparametric algorithms used for land-use–land-cover classification, there is uncertainty regarding choosing the best algorithm to measure the urban processes. In this study, several parametric (maximum likelihood) and nonparametric (support vector machine, spectral angle mapper, artificial neural network and decision tree) algorithms were used. The study was aimed at finding out the best available classification technique for land-use–land-cover classification and change detection. Landsat 8, the latest in Landsat series, and Landsat 7 and 5 freely available satellite data were used. Due to the redundancy reported for the traditional kappa-based indices, we applied modern disagreement indices to assess the accuracy of the classification process. Artificial neural network for Landsat 8 image had the highest kappa coefficient, while spectral angle mapper had the highest overall agreement (97%) and least quantity allocation error (1%). Spectral Angle Mapper gave the highest accuracy, while maximum likelihood classification gave the least for allocation and spatial disagreement indices. We found that spectral angle mapper gave the best results for land-use–land-cover change analysis in terms of least omission and commission errors (2.5% each) and highest overall agreement, whereas artificial neural network performed better in land-use–land-cover classification studies.