Spatial Outliers Research Articles

Reliability hazard characterization of wafer-level spatial metrology parameters based on LOF-KNN method

The devices with abnormal parameters often have higher early failure rates or lower reliability. The outliers identified using the wafer-level spatial metrology parameters may have high risk with the hidden reliability hazards though their parameters may meet product specifications. These devices are dangerous and are difficult to identify by existing methods. In this paper, the method was proposed to identify the wafer-level spatial metrology parameters outliers caused by defects based on LOF-KNN algorithm. First, the coordinate system of the device on the wafer was established; then the KNN (k-Nearest Neighbor) algorithm was used to characterize the spatial variation of the device, and the LOF (Local Outlier Factor) was used to characterize the local outlier of the device; finally, outliers were identified. The method was applied in the MOSFET and JBS devices cases based on the MOSFET Vth and the JBS VF wafer-level spatial metrology key parameters respectively. The outliers were obtained using the proposal algorithm. The results shown: in the MOSFET case, all non-compliant devices were identified; in additional some devices that meet product specifications but have abnormal parameters compared with their neighbors were detected. In the JBS case, where the devices all meet the product specifications, the algorithm also identified the outlier's devices with appealing abnormal parameters. Based on this work, it was demonstrated that using spatial information for outlier detection has the benefit of reducing costs and improving device reliability, which is a valuable technique.

ROBUST SPATIAL AUTOREGRESSIVE UNTUK PEMODELAN ANGKA HARAPAN HIDUP PROVINSI JAWA TIMUR

Spatial regression analysis is regression method used for type of data has a spatial effect. Spatial regression showing the presence of spatial effects on the response variable (Y) is a Spatial Autoregressive (SAR). Outlier often found in research spatial data. The outlier is called the spatial outliers. The analysis can be used to handle outliers in general is Robust Regression. There are several estimator that can be used in which the estimator Robust Regression S, M, MM and LTS. Meanwhile, Robust Regression were used to handle spatial outlier is a combination of SAR and Regression Robust method to form a new method that is Robust Spatial Autoregressive (Robust SAR). Type estimator used in this study is the S-Estimator. This study was conducted to determine the best model on a case study Life Expectancy of East Java Province. The best model is analyzed by comparing the methods of SAR and SAR Robust method. Based on the analysis results obtained MSE and Adjusted R2 values for the SAR method are 1.7521 and 55.54% while for the Robust SAR method are 0.7456 and 62.30%. The Robust SAR model has a lower MSE value and a higher Adjusted R2 when compared to the SAR model. Thus the best model for modeling the life expectancy in East Java is Robust SAR models.Keywords:Spatial Autoregressive (SAR), Robust SAR, Life expectancy

Human Development Dynamics across Districts of Indonesia: A Study of Regional Convergence and Spatial Approach

Indonesia is an archipelago country which comprises of two main parts, western and eastern regions and spread into over 500 districts. Each district has their own characteristics, especially in development aspect. Some districts have been growing faster on economic and social development, yet others still fall behind. Using Human Development Index and its components on education, health and economic variables over 2010- 2018 period, this study aims to examine convergence on the regional human growth process and investigate the speed of convergence across districts. The result reveals convergence occurred on human development process and its determinants across districts during 2010-2018. Education variables are assumed as the main contributor for boosting the speed of convergence of human development. Spatial dependences are detected among districts, followed by the spatial clusters and spatial outliers through global and local spatial autocorrelation. Applying two spatial autoregressive models, spatial autoregressive lag model (SAR) and spatial autoregressive error model (SEM), confirmed that there is significant spatial spill-overs.The speed of convergence for all variables are much declining after the inclusion of spatial lag and error model. As the policy implication, since regional inequality in term of human development is still a major issue, it will be a call for better coordination and cooperation within and between regions.

Development life expectancy model in Central Java using robust spatial regression with M-estimators

Spatial regression model is used to determine the relationship between the dependent and independent variables with spatial influence. In case only independent variables are affected, Spatial Cross Regressive (SCR) Model is formed. Spatial Autoregressive (SAR) occurs when the dependent variables are affected, while Spatial Durbin Model (SDM) exists when both variables exhibit effects. The inaccuracy of the spatial regression model can be caused by outlier observations. Removing outliers in the analysis changes the spatial effects composition on data. However, using robust spatial regression is one way of overcoming the outliers in the model. Moreover, the typical parameter coefficients, which are robust against the outliers, are estimated using M-estimator. The research develops the life expectancy model in Central Java Province through Robust-SCR, Robust-SAR, and Robust-SDM to reduce spatial outliers' effect. The model is developed based on educational, health and economic factors. According to the results, M-estimator accommodates the outliers’ existence in the spatial regression model. This is indicated by an increase in R2 value and a decrease in MSE caused by the change in the estimating coefficient parameters. In this case, Robust-SDM is the best model since it has the biggest R2 value and the smallest MSE.

Multivariate spatial analysis for the identification of criticalities and of the subtended causes in river ecosystems.

In statistics, the identification of environmental criticalities, one of the primary goals of environmental monitoring and management, translates into the detection of spatial outliers. Detected in relation to purposely defined sets of indicators, both global and local outliers are pivotal in the identification not only of the severity and spread of criticalities, but also of their nature and causes. The present research exemplifies a procedural framework to identify environmental criticalities, using two different approaches for the detection of spatial outliers in river ecosystems related to several sets of parameters (organic C, inorganic C, Ca, Co, Cr, Fe, K, Mg, Mn, N, Na, P, S, Si, V, Zn, Cl-, F-, NO3-, SO42-, chlorophyll a, chlorophyll b, pheophytin a, pheophytin b, total carotenoids, pH, and electrical conductivity), including emerging contaminants. To this end, indicator sets diagnostic for specific criticalities, derived from an empirical dataset of water quality parameters, were employed, using detection techniques based on geographically weighted principal component analysis and a modified pairwise Mahalanobis distance-based algorithm. Clear and accurate criticality scenarios were derived, highlighting both the strengths and the limitations of the proposed approach, especially in relation to the classic threshold-based methods.

GIS Based Hotspot and Cold-spot Analysis for Primary Education in India

Objectives: This study explores the status of primary education throughout the country in states and union territories (UTs). The main objectives of this study are to identify hotspots and cold-spots and delineate the cluster of primary education levels in India applying Getis-Ord Gi* statistics. Methods/analysis: The spatial clustering among the various Indian states was determined using Arc GIS cluster–outlier analysis. An area is considered statistically significant if its p-value is below 0.05. In order to define India’s spatial cluster of elementary EDI scores based on the hotspot and coldspot analysis and delineate Indian primary education clusters using Getis-Ord Gi* statistics with a fixed distance band of ArcGIS. Findings: The rate of increase of 6–10 year old (grade 1–5) enrolment in primary education rose from 42.6% in 1950–1951 to 100.8% in 2014–2015. The resulting education development index (EDI) of primary education shows a clustering of hot and cold-spots. The spatial outliers are also identified. This EDI demonstrates a geographic trend in the growth of primary education over 10 years as high or low primary education clusters and helps measure the government’s efforts. Improvements: Further this study could be enhanced to district level primary education in India. This also shows how GIS analysis works for these many more unit’s overtime. It may be an interesting analysis, where indeed the GIS and visual approach could add to our knowledge.Keywords: GIS, Hotspot Analysis, Cluster, Primary Education, Arc GIS, India

A high-resolution emissions inventory and its spatiotemporal pattern variations for heavy-duty diesel trucks in Beijing, China

Heavy-duty diesel trucks (HDDTs) cause serious pollution, and a high spatiotemporal resolution emissions inventory is a valuable assessment tool for use in quantitatively understanding the emissions mechanisms of HDDTs and scientifically developing associated emissions reduction measures. This study aims to comprehensively utilize multi-source spatiotemporal data on transportation—including fine-scale trajectories of HDDTs, road traffic conditions, and attribute data for road networks and HDDTs—supplemented by relatively mature vehicle pollution emissions models to establish a high spatiotemporal resolution emissions inventory for HDDTs in Beijing using a bottom-up approach. Spatial statistical techniques, including spatial autocorrelation, high/low clustering, and outlier analysis, are also used to explore the spatiotemporal distribution pattern of pollution emissions in the city. The results showed the following: (1) spatially, nitrogen oxide (NOx) and particulate matter (PM) emission hotspots spread from the Beijing sixth-ring roads to the fourth-ring roads from daytime to nighttime. The road segments with high emissions intensities have pronounced spatial agglomeration effects at night, but these are scattered during daytime. (2) Temporally, total HDDT NOx emissions are consistent with the traffic volume trends and are lower during major festivals. The highest NOx emissions occur at intercity highways, and this reflects the severe impact that intercity freight traffic has on air quality. The dominant HDDT NOx emissions are from vehicles belonging to the China 4 emissions standard. (3) NOx and PM emissions have a significant spatial autocorrelation and exhibit high-value clustering as a whole. (4) At different time intervals, the distribution of High-High/Low-Low clustering and outliers of NOx and PM in the road network is consistent with the spatial distribution of the pollutant emission intensity. The High-Low outlier is mainly distributed within the fourth-ring roads, and the number gradually reduces between night and day. The Low-High outlier is affected by the heterogeneous distribution of HDDTs and exhibits discontinuous distribution characteristics. Our results effectively evaluate Beijing’s emissions control measures for HDDTs and provide a scientific decision-making basis for developing targeted emission reduction strategies for HDDTs.

Which anomaly should I drill? Using spatial statistics to inform exploration in covered IOCG terranes

SummaryThis contribution presents a method for efficiently classifying geophysical anomalies and identifying regions and features that share characteristics of many known iron-oxide-copper-gold (IOCG) deposits of the Gawler Craton, and can therefore be used in drill target prioritization. Residual Bouguer gravity and reduced-to-pole total magnetic intensity grids over the Gawler Craton were transformed, generating polygon datasets representing populations of locally anomalous gravity and magnetic intensity. Taken as simple anomaly polygons, there are a very large number of features across the Gawler Craton (>39,000 TMI and >10,000 gravity). Superimposing mineral deposits over these features shows a clear spatial correlation between IOCG deposits and occurrences, and anomalies (>90% of deposits within 1,000 m of an anomaly), but leaves thousands of anomalies of varying magnitudes that cannot all be related to IOCG mineralization. Eliminating TMI and gravity anomalies with a separation of more than 1,000 m reduced the search space to ~20,000 TMI features and ~8,500 gravity features. Limiting the search to a statistically derived gravity threshold ≥0.4 mGal gravity anomalies, the exploration space is reduced to 798 gravity features with coincident TMI features within the Olympic Copper-Gold Province. The Anselin Local Morans I method was used to delineate geographic regions based upon spatial clustering of high magnitude anomalies. The spatial distribution and clustering characteristics of the gravity anomalies provide additional information and can be related to differing basement geology and deposit style. Terranes where lithologies and Cu-Au occurrences are commonly magnetite-rich show clustered high-magnitude gravity anomalies, and correlated spatially with the Mount Woods and Moonta domains within the eastern Gawler Craton. Importantly, it was found that the central, and currently most endowed, the Olympic Domain, was distinct in that it was dominated by spatial outliers (discrete high-magnitude density features).These results could be used as a starting point in developing IOCG exploration strategies, due to the high number of additional untested, spatially coincident gravity and magnetic anomalies that warrant further investigation.

Soil organic carbon distribution in relation to terrain & land use—a case study in a small watershed of Danjiangkou reservoir area, China

Understanding the responses of soil organic carbon (SOC) distribution to terrain and land use variations is important for making corresponding measures to enhance the soil quality and its carbon storage. The objective of this study was to characterize the spatial distribution of SOC, and analyze the impacts of environment factors on spatial variation of SOC along the Danjiangkou Reservoir area, China. A village size watershed was selected to investigate the SOC distribution under different terrains and land uses. We found that SOC contents followed a log-normal distribution with arithmetic and geometric means of 6.83 g kg −1 and 6.09 g kg−1, respectively, which were moderately variable (Coefficient of Variation = 49.6%) and demonstrated a moderate spatial dependence according to the nugget ratio (37.8%). The experimental variogram of SOC was best-fitted by a spherical model after the spatial outliers had been detected and subsequently eliminated. From the spatial distribution map of SOC contents, the highest SOC content occurred in the center of the watershed, which were along the main water channel or tributaries and normally had lower altitude. While,the lower values of SOC contents were located at the edge of the watershed, such as in the northwest, southwest and southeast of the area where citrus were normally planted. This spatial distribution pattern of SOC was generally consistent with the spatial structure of terrain and land use on the unique terrain of the agriculture-based reservoir area. Therefore, taking terrain and land use type into account when estimating the spatial variation of SOC pool would increase the accuracy in prediction and modeling of SOC movement at the watershed scale.

Quantile local spatial autocorrelation

This note introduces the concept of quantile local spatial autocorrelation as a special case of a local indicator of spatial association (LISA) for the situation where the variables of interest are binary. This provides additional insight into the spatial distribution of observations at the extremes of the distribution. The concept is illustrated with an analysis of local spatial clusters and outliers for health outcomes using data for 791 Chicago census tracts in 2014.

Temporal and spatial outlier detection in wireless sensor networks

Outlier detection techniques play an important role in enhancing the reliability of data communication in wireless sensor networks (WSNs). Considering the importance of outlier detection in WSNs, many outlier detection techniques have been proposed. Unfortunately, most of these techniques still have some potential limitations, that is, (a) high rate of false positives, (b) high time complexity, and (c) failure to detect outliers online. Moreover, these approaches mainly focus on either temporal outliers or spatial outliers. Therefore, this paper aims to introduce novel algorithms that successfully detect both temporal outliers and spatial outliers. Our contributions are twofold: (i) modifying the Hampel Identifier (HI) algorithm to achieve high accuracy identification rate in temporal outlier detection, (ii) combining the Gaussian process (GP) model and graph‐based outlier detection technique to improve the performance of the algorithm in spatial outlier detection. The results demonstrate that our techniques outperform the state‐of‐the‐art methods in terms of accuracy and work well with various data types.

Registration Algorithm Based on Line-Intersection-Line for Satellite Remote Sensing Images of Urban Areas

Image registration is an important step in remote sensing image processing, especially for images of urban areas, which are often used for urban planning, environmental assessment, and change detection. Urban areas have many artificial objects whose contours and edges provide abundant line features. However, the locations of line endpoints are greatly affected by large background variations. Considering that line intersections remain relatively stable and have high positioning accuracy even with large background variations, this paper proposes a high-accuracy remote sensing image registration algorithm that is based on the line-intersection-line (LIL) structure, with two line segments and their intersection. A double-rectangular local descriptor and a spatial relationship-based outlier removal strategy are designed on the basis of the LIL structure. First, the LILs are extracted based on multi-scale line segments. Second, LIL local descriptors are built with pixel gradients in the LIL neighborhood to realize initial matching. Third, the spatial relations between initial matches are described with the LIL structure and simple affine properties. Finally, the graph-based LIL outlier removal strategy is conducted and incorrect matches are eliminated step by step. The proposed algorithm is tested on simulated and real images and compared with state-of-the-art methods. The experiments prove that the proposed algorithm can achieve sub-pixel registration accuracy, high precision, and robust performance even with significant background variations.

Epidemiological and space-time analysis of Beijing Intestinal Infectious Diseases

Epidemiological and space-time analysis of Beijing Intestinal Infectious Diseases

Analyzing hemorrhagic fever with renal syndrome in Hubei Province, China: a space-time cube-based approach.

ObjectiveHemorrhagic fever with renal syndrome (HFRS), a natural–focal infectious disease caused by hantaviruses, resulted in 37 deaths between 2011 and 2015 in Hubei Province, China. HFRS outbreaks are seasonally distributed, exhibiting heterogeneity in space and time. We aimed to identify the spatial and temporal characteristics of HFRS epidemics and their probable influencing factors.MethodsWe used the space–time cube (STC) method to investigate HFRS epidemics in different space–time locations. STC can be used to visualize the trajectories of moving objects (or changing tendencies) in space and time in three dimensions. We applied space–time statistical methods, including space–time hot spot and space–time local outlier analyses, based on a calculated STC model of HFRS cases, to identify spatial and temporal hotspots and outlier distributions. We used the space–time gravity center method to reveal associations between possible factors and HFRS epidemics.ResultsIn this research, HFRS cases for each space–time location were defined by the STC model, which can present the dynamic characteristics of HFRS epidemics. The STC model delivered accurate and detailed results for the spatiotemporal patterns of HFRS epidemics.ConclusionThe methods in this paper can potentially be applied for infectious diseases with similar spatial and temporal patterns.

An Analysis of a Water Use Decoupling Index and Its Spatial Migration Characteristics Based on Extracting Trend Components: A Case Study of the Poyang Lake Basin

Water resources and their utilization perform a critical role in sustainable development. A full comprehension of the decoupling relationship between water consumption and economic development is a prerequisite for sustainable water resource management. This thesis developed a decoupling index analysis model based on Hodrick–Prescott filtering; analyzed the spatial aggregation characteristics of Gross Domestic Product (GDP), water consumption, and the decoupling index by the Global and Local Moran’s Index; and calculated the spatial gravity migration characteristics of GDP, water consumption, and the decoupling index. A case study in the Poyang Lake basin was selected to analyze the relationship between water and the economy. The results indicated that decoupling status was steadier after extracting trend components. The decoupling index exhibited spatial outlier characteristics. The spatial gravity center migration directions of GDP and water consumption were opposite. Furthermore, the Poyang Lake basin was in a weak decoupling status, and its water use pattern was sustainable to a certain extent.

Geospatial Correlation of Amyopathic Dermatomyositis With Fixed Sources of Airborne Pollution: A Retrospective Cohort Study.

Objective: Dermatomyositis (DM) may result from exogenous triggers, including airborne pollutants, in genetically susceptible individuals. The United States Environmental Protection Agency's 2011 National Air Toxics Assessment (NATA) models health risks associated with airborne emissions, available by ZIP code tabulation area (ZCTA). Important contributors include point (fixed), on-road, and secondary sources. The objective of this study was to investigate the geospatial distributions of DM and subtypes, classic DM (CDM) and clinically amyopathic DM (CADM), and their associations with airborne pollutants.Methods: This retrospective cohort study identified 642 adult DM patients from 336 unique ZCTAs. GeoDa v.1.10 was used to calculate global and local Moran's indices and generate local indicator of spatial autocorrelation (LISA) maps. All Moran's indices and LISA maps were permuted 999 times.Results: Univariate global Moran's indices for DM, CDM, and CADM prevalence were not significant, but LISA maps demonstrated differential local spatial clustering and outliers. CADM prevalence correlated with point sources (bivariate global Moran's index 0.071, pseudo-p = 0.018), in contrast to CDM (−0.0053, pseudo-p = 0.46). Bivariate global Moran's indices for DM, CDM, and CADM prevalence did not correlate with other airborne toxics, but bivariate LISA maps revealed local spatial clustering and outliers.Conclusion: Prevalence of CADM, but not CDM, is geospatially correlated with fixed sources of airborne emissions. This effect is small but significant and may support the hypothesis that triggering exposures influence disease phenotype. Important limitations are NATA data and ZCTA population estimates were collected from 2011 and ZCTA of residence may not have been where patients had greatest airborne pollutant exposure.

Errors in Time-Series Remote Sensing and an Open Access Application for Detecting and Visualizing Spatial Data Outliers Using Google Earth Engine

Remotely sensed measures of productivity are frequently used to characterize global agriculture and vegetated ecosystems, and are often downscaled to describe local, remote areas where finer spatial and temporal resolution data are regularly unavailable. While data errors may propagate throughout any analytical procedure, those that are missed during delivery and preliminary data mining require more attention. Here, a collection of formerly and presently available global remote sensing products are compiled to demonstrate the temporal and geographic breadth of remote sensing uncertainty. Vegetation productivity measures are invaluable for monitoring global health, but erroneous estimates that go unrecognized may result in serious policy mistakes. It is eminently clear that generalizable and accessible a priori methods for anomaly detection are lacking and urgently needed so that data errors are recognized before public delivery and before widespread use. Simple yet effective statistics such as the modified Z-score, Tukey's outliers, and Geary's C are leveraged here to identify, locate, and visualize the types of outliers that remote sensing data users may elect to omit or correct. Contributing to the growing ensemble of Google Earth Engine methodologies, we propose this generalizable method of detecting spatial outliers for remote sensing error management by users across scientific domains.

The application of Local Moran's I to identify spatial clusters and hot spots of Pb, Mo and Ti in urban soils of Yerevan

Soils are the main sink of pollutants in the terrestrial environment, reflecting the historical fingerprint of urban development. To provide scientifically grounded urban management and reclamation strategies for polluted sites, the spatial patterns of pollutants should be studied first by drawing a special focus on the identification of pollution hot spots. In this paper, Pb, Mo and Ti contents in Yerevan urban soils determined by the combination of X-ray fluorescence and atomic absorption spectrometry were studied using the Local Moran's I index (LMI). These elements (Pb, Mo and Ti) have different sources and a number of distance bands (650, 800, 1000, 1500, 2000, 3000, 4000, 5000, 6000, 7000, 8000, 9000, 10,000 m and 15,000 m) were tested. This allows to reveal the optimal distance used for the identification of statistically significant spatial clusters and spatial outliers. According to the results the use of the raw data had to be excluded for elements that demonstrated highly skewed distribution whereas transformed data allowed to identify larger areas of high and low value spatial clusters. Moreover, the results of the different distance bands showed that the high and low spatial clusters became stable near the optimal distance band of 7000 m in the study area. The latter constituted the 1/3 of the maximum distance between all sample pairs. At this distance, Ti LMI results reflected the natural local geology composition whereas Mo distribution was in line with the known sources of industrial pollution (“Plant of Pure Iron”). In the case of Pb, the high-value spatial clusters were in line with the built-up urban areas and spatially correlated with the prevailing northeast winds. This study confirmed the applicability of LMI in the identification of spatial clusters and outliers at the optimal distance band which might be considered also as a break point of spatial clusters expansion and a guidance for spatial boundary of possible remediation and pollution reduction applications.

Protocol for automating error removal from yield maps

Yield mapping is one of the most widely used precision farming technologies. However, the value of the maps can be compromised by the presence of systematic and random errors in raw within field data. In this paper, an automated method to clean yield maps is proposed so as to ensure the quality of further data processing and management decisions. First, data were screened by filtering null and edge yield values as well global outliers. Second, spatial outliers or local defective observations were deleted. The local Moran’s index of spatial autocorrelation and Moran’s plot were used as tool to identify the spatial outliers. The protocol to filter out global and local outliers was evaluated on 595 real yield datasets from different grain crops. Significant improvements in the distribution and spatial structure of yield datasets was found. Approximately 30% of the dataset size was removed from each monitor dataset, with one third of the removal occurring during filtering of spatial outliers. The automation of null, edge yield values and the removal of global outliers improved yield distributions, whereas the cleaning of local outliers impacted the yield spatial structure for all yield maps and crops. The algorithm proposed to clean yield maps is easy to apply for preprocessing the growing number of available yield maps.

Bike sharing systems: a new incentive rebalancing method based on spatial outliers detection

Since its launch, Velib' (the Bike Sharing System – BSS – in Paris) has emerged in the Parisian landscape and has been a model for similar systems in many cities. A major problem with BSS is the stations' heterogeneity caused by the attractivity of some stations located in particular areas. In this paper, we focus on spatial outliers defined as stations having a behaviour significantly different from their neighbouring stations. First, we propose an improved version of Moran scatterplot to exploit the similarity between neighbours, and we test it on a real dataset issued from Velib' system to identify outliers. Then, we design a new method that globally improves the resources' availability in bike stations by adapting the users' trips to the resources' availability. Results show that with a partial collaboration of the users or a limitation to the rush hours, the proposed method enhances significantly the resources' availability in Velib' system.