High Outliers Research Articles

Outlier ophthalmologists in the treatment of neovascular age-related macular degeneration with intravitreal therapy

BackgroundTo compare individual ophthalmologists grouped as outliers or non-outliers based on the mean 12-month visual acuity (VA) outcomes for their patients with neovascular age-related macular degeneration (nAMD).MethodsThis prospectively designed database...

A robust correspondence-based registration method for large-scale outdoor point cloud

ABSTRACT Large-scale outdoor point cloud registration is essential for the 3D reconstruction of outdoor scenes. Its central objective is to achieve accurate point cloud registration by determining accurate spatial transformation parameters. While feature-based methods eliminate the need for initial position estimation, they encounter challenges in handling high outlier rates. Therefore, a method capable of effectively managing outliers is crucial for enhancing the efficiency and accuracy of large-scale outdoor point cloud registration. This paper introduces the maximal clique with adaptive voting (MCAV) method, which leverages graph-based inlier compatibility to optimize potential matches. MCAV employs adaptive parameter voting (APV) to enhance computational efficiency, demonstrating significant speedup characteristics in datasets with a significant number of inliers. To further reduce outliers in potential matches, we integrate Black-Rangarajan Duality (BRD) and graduated non-convexity (GNC) into the truncated least squares (TLS) framework (BG-TLS). Accordingly, we propose the efficient BG-TLS (EBG-TLS) method for computing the registration model. Comparative analyses with traditional and deep learning-based methods across various real-world environments demonstrate that the proposed method outperforms existing algorithms in terms of rotation error, translation error, and efficiency, particularly in complex, high-noise settings. This method finds broad applications in geospatial mapping and surveying, autonomous navigation, and environmental monitoring.

Abstract A057: Superfund sites and cancer risk: Cluster analysis in Puerto Rican municipalities between 2010 and 2020

Abstract Introduction: Living within proximity to hazardous waste (superfund) sites can pose an increased risk of exposure to chemicals of concern, including cancerogenic substances. Per- square mile, Puerto Rico has a high density of contaminated sites, including the designation of 19 superfund sites by the Environmental Protection Agency across the archipelago. Understanding how living within proximity to superfund sites can help recognize communities that are at greater health risks, such as increased cancer risks. Objectives: Identify clusters and outliers for all types of cancer, to assess ecological risk, by the presence of superfund sites across Puerto Rico. Methods: All types of cancer were recorded and provided by the Puerto Rico Central Cancer Registry for the period of 2010-2020. This data was summarized by municipalities (n=78). Cancer cases in each county were aggregated and divided by population in 2020 to show prevalence of cases. Using geographical information systems (ArcGIS), a Local Moran’s I was run to identify potential municipalities that have clusters; location with similar high (hot spots) or low values (cold spots) and outliers; locations with dissimilar values from their surrounding neighbors. A layer representing municipalities with superfund sites was also applied to the map to see if there was a relationship between living in a superfund site and cancer risk. This study was approved by UPRCCC IRB Protocol #2023-11-124. Results: Between 2010 and 2020, a total of 114,560 diagnosis occurred for all types of cancer in PR, from which 48.06% accounted for females and 51.93% for males, with breast and prostate as the primary cancer sites, respectively. Our analyses showed one hotspot and three cold spots with statistically significant differences from their neighboring municipalities. One hotspot was in Canovanas (p=0.03) which has a superfund site related to pharmaceutical chemicals. The three coldspots were in Cataño (p=0.04), Trujillo Alto (p=0.04), and Guánica (p=0.03). Additionally, four high and low outliers were identified. Across the municipalities, in the lower outliers Vega Alta (p=0.04), Corozal (p=0.002) are the only ones with current Superfund sites. Cabo Rojo (p=0.002) and San German (p=0.006) are included in the list of higher outliers, which are places with superfund sites. However, Hormigueros (p=0.03), and Arroyo (p=0.03), are also in this category. Conclusion: The results identify the key cancer clusters and outliers across municipalities in comparison to neighboring municipalities on the island. Results show municipalities with and without Superfund sites as part of the higher outliers list, which additional research could help establish other contributing factors. Further analysis is needed to consider what factors affect the clusters such as age range, type of cancer, and presence of other environmental contaminants in the area. Acknowledgements: This research was supported by award 1R25CA240120-01 (Cancer Prevention & Control Research Training Program) from the National Cancer Institute. Citation Format: Sailing JT Tang, Nancy R. Cardona, Ana P. Ortiz. Superfund sites and cancer risk: Cluster analysis in Puerto Rican municipalities between 2010 and 2020 [abstract]. In: Proceedings of the 17th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2024 Sep 21-24; Los Angeles, CA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2024;33(9 Suppl):Abstract nr A057.

Hospital variation in adoption of balanced transfusion practices among injured patients requiring blood transfusions

BackgroundThis study sought to measure hospital variability in adoption of balanced transfusion following the Pragmatic, Randomized Optimal Platelet and Plasma Ratios (PROPPR) guidelines. We hypothesized hospital adoption rates of balanced transfusion would be low, and vary significantly among hospitals after controlling for patient, injury and hospital characteristics. Study Design and MethodsThis was an observational cohort study of injured adult patients (≥16 years) in Trauma Quality Improvement Program hospitals 2016–2021. Inclusion criteria were hypotensive patients receiving one transfusion of packed red blood cells, fresh frozen plasma, platelets, or cryoprecipitate. Balanced transfusion was defined as ≥1 ratio of plasma to packed red blood cells or platelets to packed red blood cells or whole blood use at 4 hours. Hierarchical multivariable logistic regression quantified residual hospital-level variability in balanced transfusion rates after adjusting for patient and hospital characteristics. ResultsAmong 172,457 injured patients who received transfusions, 30,386 (17.6%) underwent balanced transfusion. Patient-level balanced transfusion rates were 11% in 2016, rose to 14.0% in 2019, and jumped up once whole blood transfusions were measured to 24.0% in 2020 and to 25.9% in 2021. Approximately 26% of the variability in balanced transfusion rates was attributable to the hospital. Verified level I hospitals had a 2.09 increased adjusted odds of balanced transfusion (95% CI 1.88–2.21) compared to nonverified hospitals. University teaching status had a 1.29 increased adjusted odds of balanced transfusion (95% CI 1.08–1.54) compared with community hospitals. Overall, 150 (23.5%) hospitals were high outliers (high performing) in balanced transfusion adoption and 124 (19.4%) hospitals were low outliers. ConclusionThere was significant variability in hospital adoption of balanced transfusion.

Variation in Risk-Adjusted Ventilator-Associated Pneumonia Days Within a Quality Collaborative

IntroductionVentilator-associated pneumonia (VAP) is associated with increased mortality, prolonged mechanical ventilation, and longer intensive care unit stays. The rate of VAP (VAPs per 1000 ventilator days) within a hospital is an important quality metric. Despite adoption of preventative strategies, rates of VAP in injured patients remain high in trauma centers. Here, we report variation in risk-adjusted VAP rates within a statewide quality collaborative. MethodsUsing Michigan Trauma Quality Improvement Program data from 35 American College of Surgeons–verified Level I and Level II trauma centers between November 1, 2020 and January 31, 2023, a patient-level Poisson model was created to evaluate the risk-adjusted rate of VAP across institutions given the number of ventilator days, adjusting for injury severity, physiologic parameters, and comorbid conditions. Patient-level model results were summed to create center-level estimates. We performed observed-to-expected adjustments to calculate each center's risk-adjusted VAP days and flagged outliers as hospitals whose confidence intervals lay above or below the overall mean. ResultsWe identified 538 VAP occurrences among a total of 33,038 ventilator days within the collaborative, with an overall mean of 16.3 VAPs per 1000 ventilator days. We found wide variation in risk-adjusted rates of VAP, ranging from 0 (0-8.9) to 33.0 (14.4-65.1) VAPs per 1000 d. Several hospitals were identified as high or low outliers. ConclusionsThere exists significant variation in the rate of VAP among trauma centers. Investigation of practices and factors influencing the differences between low and high outlier institutions may yield information to reduce variation and improve outcomes.

Calculation and Feedback of Risk-Adjusted Antibiotic Days as a Process Measure in a Statewide Trauma Collaborative.

Optimization of antibiotic stewardship requires determining appropriate antibiotic treatment and duration of use. Our current method of identifying infectious complications alone does not attempt to measure the resources actually utilized to treat infections in patients. We sought to develop a method accounting for treatment of infections and length of antibiotic administration to allow benchmarking of trauma hospitals with regard to days of antibiotic use. Using trauma quality collaborative data from 35 American College of Surgeons (ACS)-verified level I and level II trauma centers between November 1, 2020, and January 31, 2023, a two-part model was created to account for (1) the odds of any antibiotic use, using logistic regression; and (2) the duration of usage, using negative binomial distribution. We adjusted for injury severity, presence/type of infection (eg, ventilator-acquired pneumonia), infectious complications, and comorbid conditions. We performed observed-to-expected adjustments to calculate each center's risk-adjusted antibiotic days, bootstrapped Observed/Expected (O/E) ratios to create confidence intervals, and flagged potential high or low outliers as hospitals whose confidence intervals lay above or below the overall mean. The mean antibiotic treatment days was 1.98°days with a total of 88,403 treatment days. A wide variation existed in risk-adjusted antibiotic treatment days (.76°days to 2.69°days). Several hospitals were identified as low (9 centers) or high (6 centers) outliers. There exists a wide variation in the duration of risk-adjusted antibiotic use amongst trauma centers. Further study is needed to address the underlying cause of variation and for improved antibiotic stewardship.

Using implementation science to decrease variation and high opioid administration in a surgical ICU.

High doses and prolonged duration of opioids are associated with tolerance, dependence, and increased mortality. Unfortunately, despite recent efforts to curb outpatient opioid prescribing because of the ongoing epidemic, utilization remains high in the intensive care setting, with intubated patients commonly receiving infusions with a potency much higher than doses required to achieve pain control. We attempted to use implementation science techniques to monitor and reduce excessive opioid prescribing in ventilated patients in our Surgical ICU. We conducted a prospective study investigating opioid administration in a closed SICU at an academic medical center over 18 months. Commonly accepted conversions were used to aggregate daily patient opioid use. Patients with a history of chronic opioid use and those being treated with an ICP monitor/drain, neuromuscular blocker, or ECMO were excluded. If the patient spent a portion of a day on a ventilator, that day's total was included in the "vent group." MMEs per patient were collected for each patient and assigned to the on-call intensivist. Intensivists were blinded to the data for the first seven months. They were then provided with academic detailing followed by audit & feedback over the subsequent 11 months, demonstrating how opioid utilization during their time in the SICU compared to the unit average and a blinded list of the other attendings. Student's T-tests were performed to compare opioid utilization before and after initiation of academic detailing and audit & feedback. Opioid utilization in patients on a ventilator decreased by 20.1% during the feedback period, including less variation among all intensivists and a 30.9% reduction by the highest prescribers. Implementation science approaches can effectively reduce variation in opioid prescribing, especially for high outliers in a SICU. These interventions may reduce the risks associated with prolonged use of high-dose opioids. Prospective pre-post-intervention, Level II.

Non-neurological factors associated with serum neurofilament levels in the United States population.

To model interdependencies of serum neurofilament light chain (sNfL), a clinically useful biomarker of axonal injury in neurological diseases, with demographic, anthropometric, physiological, and disease biomarkers in the United States population. sNfL and 80 biomarkers were obtained from the National Health and Nutrition Examination Survey (n = 2071, age: 20-75 years). Body habitus and composition, electrolytes, blood cell, metabolic, liver, and kidney function biomarkers, and common diseases were assessed with weighted regression adjusted for age, sex, and race/ethnicity. Salient biomarkers were modeled with ensemble learning; a Bayesian network structure was obtained for interdependencies. Age was strongly associated with sNfL. sNfL levels were 13% higher in men versus women. Mexican Americans had 18.5% lower sNfL versus Non-Hispanic Whites. sNfL was similar in pregnant versus nonpregnant women. Lymphocyte, and neutrophil numbers, and phosphorus, and chloride levels were associated with sNfL. Multiple liver function (e.g., albumin and gamma-glutamyltransferase), renal function (e.g., creatinine and urea), and carbohydrate/lipid metabolism markers (e.g., glucose and triglycerides) were associated with sNfL. A 50% greater creatinine was associated with 26.8% greater sNfL. Diabetes, kidney disease, congestive heart failure, and stroke were associated with sNfL. The ensemble learning algorithm predicted high sNfL outliers with 5.06%-9.16% test error. Bayesian network modeling indicated sNfL had neighbor dependencies with age, creatinine, albumin, and chloride. sNfL is associated with age, kidney and liver function, diabetes, blood cell subsets, and electrolytes. sNfL may be a useful biomarker for biological age of the whole body and major organ systems including the brain.

Multi-Echo Complex Quantitative Susceptibility Mapping and Quantitative Blood Oxygen Level-Dependent Magnitude (mcQSM + qBOLD or mcQQ) for Oxygen Extraction Fraction (OEF) Mapping.

Oxygen extraction fraction (OEF), the fraction of oxygen that tissue extracts from blood, is an essential biomarker used to directly assess tissue viability and function in neurologic disorders. In ischemic stroke, for example, increased OEF can indicate the presence of penumbra-tissue with low perfusion yet intact cellular integrity-making it a primary therapeutic target. However, practical OEF mapping methods are not currently available in clinical settings, owing to the impractical data acquisitions in positron emission tomography (PET) and the limitations of existing MRI techniques. Recently, a novel MRI-based OEF mapping technique, termed QQ, was proposed. It shows high potential for clinical use by utilizing a routine sequence and removing the need for impractical multiple gas inhalations. However, QQ relies on the assumptions of Gaussian noise in susceptibility and multi-echo gradient echo (mGRE) magnitude signals for OEF estimation. This assumption is unreliable in low signal-to-noise ratio (SNR) regions like disease-related lesions, risking inaccurate OEF estimation and potentially impacting clinical decisions. Addressing this, our study presents a novel multi-echo complex QQ (mcQQ) that models realistic Gaussian noise in mGRE complex signals. We implemented mcQQ using a deep learning framework (mcQQ-NET) and compared it with the existing QQ-NET in simulations, ischemic stroke patients, and healthy subjects, using identical training and testing datasets and schemes. In simulations, mcQQ-NET provided more accurate OEF than QQ-NET. In the subacute stroke patients, mcQQ-NET showed a lower average OEF ratio in lesions relative to unaffected contralateral normal tissue than QQ-NET. In the healthy subjects, mcQQ-NET provided uniform OEF maps, similar to QQ-NET, but without unrealistically high OEF outliers in areas of low SNR, such as SNR ≤ 15 (dB). Therefore, mcQQ-NET improves OEF accuracy by more accurately reflecting realistic Gaussian noise in complex mGRE signals. Its enhanced sensitivity to OEF abnormalities, based on more realistic biophysics modeling, suggests that mcQQ-NET has potential for investigating tissue variability in neurologic disorders.

Implementation of the Caprini risk assessment model (RAM) in surgical patients to decrease postsurgical venous thromboembolism and enoxaparin prescription at hospital discharge

Venous thromboembolism (VTE), including deep venous thrombosis (DVT) and pulmonary embolism (PE), has been an increasingly common post-surgical complication for surgical patients. In the United States, VTE has become a leading cause of preventable hospital death with more than half occurring after discharge and are directly linked to a recent (within 30 days) hospitalization or surgery [1]. In large, hospital-associated/acquired VTE (HA-VTE) are preventable through measures such as the use of risk stratification tools and chemoprophylaxis. The project institution, a community, academic, medical center, for multiple years has consistently remained a high outlier for postoperative VTE. Also, the choice of VTE chemoprophylaxis in surgical patients at the time of discharge depended on, and varied between, the individual prescribing physician. The goal was to implement and determine the efficacy of a standardized intervention tool, the Caprini risk assessment model (RAM), for reducing postoperative VTE complications and its influence on the physician's prescription of enoxaparin at discharge. Results: Risk assessment scoring pre-operatively increased from 0% baseline to 26.3% at Plan-Do-Study-Act (PDSA) cycle 1 and demonstrated a statistically significant change (p-value = 0.006). Risk assessment scoring pre-operatively was 42.9% by PDSA cycle 2 but was not statistically significantly different from PDSA cycle 1. Risk assessment scoring post-operatively (for eligible patients) remained the same throughout all three cycles at 0%. Appropriate prescription of anticoagulation declined from baseline (12.5%) to PDSA cycle 1 (0%), and improved at PDSA cycle 2 (33.3%), however no differences were significant (p-value 0.302). The National Surgical Quality Improvement Project (NSQIP) database showed a decline in VTE occurrences at the projects institution from baseline (1.02%, 6 occurrences, 2021) to PDSA cycle 2 (0.92%, 4 occurrences, 2022) when compared to the national benchmark (1.0%) for the first time since 2018. Given the significant national problem HA-VTE pose to the public, and the rise in occurrences, this quality improvement (QI) project is clinically relevant.

T distribution-based robust semiparametric mixture regression model

Semiparametric mixture of regression (SMR) models provide a popular and flexible framework for modeling heterogeneous data that violates some of the parametric assumptions assumed in traditional finite mixture of regressions models. The majority of applications of SMR models assume normality for their error terms. As is well known, Gaussian distribution is sensitive to outliers or heavy-tailed distribution. In this article, we propose a more robust approach of SMR by modeling the error distribution as t distributions. By combining the EM algorithm and kernel density estimator, two algorithms are proposed to fit the robust SMR models and are proven to monotonically increase the likelihood function. We further investigate a modified version based on trimming for high leverage outliers. In real applications, a data adaptive mechanism is discussed to choose the degrees of freedom. A simulation study and three real data analyses show the superiority of the new methodology.

Investigating trends in traffic count data utilizing exploratory data analysis and regression analysis

To decipher and understand data collected for any model or decision-making, a statistical investigation into a dataset is necessary to generate every unearthed information in the data through visual means or a summary of the dataset’s patterns. Statistical analysis is utilized to test: normality, randomness, outliers, the relationship between the dependent and the independent variables, and possible temporal and spatial spread in the dataset. In this study, the annual average daily traffic (AADT) dataset obtained from Montana, Minnesota, and Washington were subjected to statistical investigations using descriptive statistics, regression analysis, and hypothesis testing to understand the data patterns. The results generated from exploring the various techniques indicated that the data acquired were generally random. However, the data’s orderliness is not random. The data values had nearly no relationship with the collection locations, thus requiring a model that can generate a precise fit for better decision-making. Data skewness varied from slight to high. Therefore, data ranged from approximating normality and nonnormality with high outliers.

Estimating pitting descriptors of 316 L stainless steel by machine learning and statistical analysis

A hybrid rule-based/ML approach using linear regression and artificial neural networks (ANNs) determined pitting corrosion descriptors from high-throughput data obtained with Scanning Electrochemical Cell Microscopy (SECCM) on 316 L stainless steel. Non-parametric density estimation determined the central tendencies of the Epit/log(jpit) and Epass/log(jpass) distributions. Descriptors estimated using conditional mean or median curves were compared to their central tendency values, with the conditional medians providing more accurate results. Due to their lower sensitivity to high outliers, the conditional medians were more robust representations of the log(j) vs. E distributions. An observed trend of passive range shortening with increasing testing aggressiveness was attributed to delayed stabilisation of the passive film, rather than early passivity breakdown.

Reducing Unplanned Intubations in the Neonatal Intensive Care Unit After Children's Surgery: A Quality Improvement Project

BackgroundUnplanned intubation following children's surgery is associated with increased postoperative mortality. In response to being a National Surgical Quality Improvement Program - Pediatric (NSQIP-P) high outlier for postoperative unplanned intubation, we aimed to reduce postoperative unplanned intubation events by 25% in one year. Methods/InterventionA multidisciplinary team of stakeholders was assembled in 2018. Most unplanned intubation events occurred in the neonatal intensive care unit (NICU). Based on apparent causes of unplanned intubations identified in case reviews, an extubation readiness checklist and a postoperative pain management guideline emphasizing non-opioid analgesics were implemented for NICU patients in September 2019. Postoperative unplanned intubation events were tracked prospectively and evaluated using quality improvement statistical process control methods. ResultsUnplanned intubations in the NICU decreased from 0.27 to 0.07 events per patient in the post-intervention group (September 2019–June 2022, n = 145) compared to the pre-intervention group (January 2016–August 2019, n = 200), representing a 76% reduction. Postoperative opioid administration decreased significantly, while acetaminophen usage increased significantly over time. Balancing measures of postoperative pneumonia rate (1.5% vs 0.0%, p = 0.267) and median hospital length of stay [40 (IQR 51) days vs 27 (IQR 60), p = 0.124] were not different between cohorts. The 30-day mortality rate for postoperative patients in the NICU significantly declined [6.5% (n = 13) vs 0.7% (n = 1), p < 0.001]. ConclusionsPostoperative unplanned intubation rates for NICU patients decreased following a quality improvement effort focused on opioid stewardship and extubation readiness. Type of StudyProspective Quality Improvement. Level of EvidenceLevel III.

Influence diagnostics in geographically weighted ridge regression

The presence of influential observations affect the accuracy of the model results. Similarly, the multicollinearity is the other problem that affect accuracy. Literature has shown that both the presence of influential observations and multicollinearity can arise in the model at the same time. This situation is also encountered in the geographically weighted regression. This study introduces the case-deletion diagnostics for the ridge estimator in geographically weighted regression model. To identify the influential observations, Cook’s Distance and Pena’s Statistics are adapted for the geographically weighted ridge regression model. The proposed diagnostics have been applied through a well-known real data. The influential observations have been detected by Cook’s Distance and Pena’s Statistics plots in real data. In addition, the simulated data has been used to reveal the performance of the proposed diagnostics. The results of the simulation showed that Pena’s Statistics has been very good at detecting high leverage outliers in the large data sets, especially.

Halfway to Automated Feeding of Chinese Hamster Ovary Cells.

This paper presents a comprehensive study on the development of models and soft sensors required for the implementation of the automated bioreactor feeding of Chinese hamster ovary (CHO) cells using Raman spectroscopy and chemometric methods. This study integrates various methods, such as partial least squares regression and variable importance in projection and competitive adaptive reweighted sampling, and highlights their effectiveness in overcoming challenges such as high dimensionality, multicollinearity and outlier detection in Raman spectra. This paper emphasizes the importance of data preprocessing and the relationship between independent and dependent variables in model construction. It also describes the development of a simulation environment whose core is a model of CHO cell kinetics. The latter allows the development of advanced control algorithms for nutrient dosing and the observation of the effects of different parameters on the growth and productivity of CHO cells. All developed models were validated and demonstrated to have a high robustness and predictive accuracy, which were reflected in a 40% reduction in the root mean square error compared to established methods. The results of this study provide valuable insights into the practical application of these methods in the field of monitoring and automated cell feeding and make an important contribution to the further development of process analytical technology in the bioprocess industry.

Automated Detection of Patients at High Risk of Polypharmacy including Anticholinergic and Sedative Medications.

Ensuring that medicines are prescribed safely is fundamental to the role of healthcare professionals who need to be vigilant about the risks associated with drugs and their interactions with other medicines (polypharmacy). One aspect of preventative healthcare is to use artificial intelligence to identify patients at risk using big data analytics. This will improve patient outcomes by enabling pre-emptive changes to medication on the identified cohort before symptoms present. This paper presents a mean-shift clustering technique used to identify groups of patients at the highest risk of polypharmacy. A weighted anticholinergic risk score and a weighted drug interaction risk score were calculated for each of 300,000 patient records registered with a major regional UK-based healthcare provider. The two measures were input into the mean-shift clustering algorithm and this grouped patients into clusters reflecting different levels of polypharmaceutical risk. Firstly, the results showed that, for most of the data, the average scores are not correlated and, secondly, the high risk outliers have high scores for one measure but not for both. These suggest that any systematic recognition of high-risk groups should consider both anticholinergic and drug-drug interaction risks to avoid missing high-risk patients. The technique was implemented in a healthcare management system and easily and automatically identifies groups at risk far faster than the manual inspection of patient records. This is much less labour-intensive for healthcare professionals who can focus their assessment only on patients within the high-risk group(s), enabling more timely clinical interventions where necessary.

دمج البوتستراب الاحتمالي الموزون مع طريقة لارس الحصينة لاختيار المُتغيّرات في أنموذج الانحدار الخطّي بوجود مشكلتي الأبعاد العالية والقيم الشاذّة.

In this research, a new algorithm was proposed to select the important variables in the regression model with the presence of two problems of high dimensions and outliers by employing and integrating the weighted bootstrap probability - Robust Least Angle Regression Selecting (WBP-LARS) and comparing it with another selection method. It is a method of impregnable Lars based on the regular bootstrap method, known as (B-LARS), empirically simulated and applied, based on real data related to the market value of some private banks in the stock market for the period 2010-2017. The comparison in the simulation included two cases for the required number of explanatory variables Choosing (K = 5, K = 7) as well as two cases when (n&gt;P) (n&lt;P) and sample sizes (50, 70, 20, 26) with a correlation value of 0.95 and with different contamination rates α = (0.05, 0.10, 0.15) The research concluded with conclusions, the most important of which is determining the number of important variables between the numbers 7-10, the preference of the (WBP-LARS) method over the (B-LARS) method when (n&gt;P), while a slight preference for the (B-LARS) method appeared over the proposed method when ( n&lt;P) and the sample size is far from the total number of variables in the model, but the efficiency converges whenever the sample size is very close to the number of variables and can be relied upon in the selection process for the variables in this case

AutoOD: Automatic Outlier Detection

Outlier detection is critical in real world. Due to the existence of many outlier detection techniques which often return different results for the same data set, the users have to address the problem of determining which among these techniques is the best suited for their task and tune its parameters. This is particularly challenging in the unsupervised setting, where no labels are available for cross-validation needed for such method and parameter optimization. In this work, we propose AutoOD which uses the existing unsupervised detection techniques to automatically produce high quality outliers without any human tuning. AutoOD's fundamentally new strategy unifies the merits of unsupervised outlier detection and supervised classification within one integrated solution. It automatically tests a diverse set of unsupervised outlier detectors on a target data set, extracts useful signals from their combined detection results to reliably capture key differences between outliers and inliers. It then uses these signals to produce a "custom outlier classifier" to classify outliers, with its accuracy comparable to supervised outlier classification models trained with ground truth labels - without having access to the much needed labels. On a diverse set of benchmark outlier detection datasets, AutoOD consistently outperforms the best unsupervised outlier detector selected from hundreds of detectors. It also outperforms other tuning-free approaches from 12 to 97 points (out of 100) in the F-1 score.

The Spatiotemporal Pattern and Driving Mechanism of Urban Sprawl in China’s Counties

Cities in China do not constitute a few global metropolises, but are characterized by heterogeneity. Studying counties can give us a comprehensive picture of urban sprawl in China. This study measured the sprawl index of 1880 counties in China from 2005 to 2020 for the first time and then revealed the evolution of their spatiotemporal characteristics and driving mechanisms. The results revealed the following. (1) China’s counties had a noticeable sprawling trend from 2005 to 2020, and their evolutionary process was characterized by spatiotemporal heterogeneity. (2) From 2005 to 2020, the counties’ sprawl gradually evolved into a spatial distribution pattern of high in the east and low in the west. The spatial distribution of sprawl in county and municipal districts had the characteristics of an interlocking distribution. (3) High–high cluster areas of CSI are mainly distributed in plains, and hilly, basin, and plateau areas tend to be low–low cluster areas. High–low outliers were distributed in a “point–line” pattern along the railroad lines and a cluster pattern near railroad intersections and central cities. Low–high outliers had the trend of encircling the high–high cluster areas. (4) The coefficient of the natural drivers was higher but tended to decrease, while the coefficient of economic and spatial drivers was lower but gradually increased. This study is the first to refine the study of urban sprawl to the county scale, which provides a reference for decision making to optimize the spatial structure of counties and thus promote high-quality development.