Social Network Forensics Analysis Model Based on Network Representation Learning.

The rapid evolution of computer technology and social networks has led to massive data generation through interpersonal communications, necessitating improved methods for information mining and relational analysis in areas such as criminal activity. This paper introduces a Social Network Forensic Analysis model that employs network representation learning to identify and analyze key figures within criminal networks, including leadership structures. The model incorporates traditional web forensics and community algorithms, utilizing concepts such as centrality and similarity measures and integrating the Deepwalk, Line, and Node2vec algorithms to map criminal networks into vector spaces. This maintains node features and structural information that are crucial for the relational analysis. The model refines node relationships through modified random walk sampling, using BFS and DFS, and employs a Continuous Bag-of-Words with Hierarchical Softmax for node vectorization, optimizing the value distribution via the Huffman tree. Hierarchical clustering and distance measures (cosine and Euclidean) were used to identify the key nodes and establish a hierarchy of influence. The findings demonstrate the effectiveness of the model in accurately vectorizing nodes, enhancing inter-node relationship precision, and optimizing clustering, thereby advancing the tools for combating complex criminal networks.

The determinants of COVID-19 case reporting across Africa.

Background: According to study on the under-estimation of COVID-19 cases in African countries, the average daily case reporting rate was only 5.37% in the initial phase of the outbreak when there was little or no control measures. In this work, we aimed to identify the determinants of the case reporting and classify the African countries using the case reporting rates and the significant determinants. Methods: We used the COVID-19 daily case reporting rate estimated in the previous paper for 54 African countries as the response variable and 34 variables from demographics, socioeconomic, religion, education, and public health categories as the predictors. We adopted a generalized additive model with cubic spline for continuous predictors and linear relationship for categorical predictors to identify the significant covariates. In addition, we performed Hierarchical Clustering on Principal Components (HCPC) analysis on the reporting rates and significant continuous covariates of all countries. Results: 21 covariates were identified as significantly associated with COVID-19 case detection: total population, urban population, median age, life expectancy, GDP, democracy index, corruption, voice accountability, social media, internet filtering, air transport, human development index, literacy, Islam population, number of physicians, number of nurses, global health security, malaria incidence, diabetes incidence, lower respiratory and cardiovascular diseases prevalence. HCPC resulted in three major clusters for the 54 African countries: northern, southern and central essentially, with the northern having the best early case detection, followed by the southern and the central. Conclusion: Overall, northern and southern Africa had better early COVID-19 case identification compared to the central. There are a number of demographics, socioeconomic, public health factors that exhibited significant association with the early case detection.

Genetic diversity and population structure of Uganda cassava germplasm.

Cassava (Manihot esculenta Crantz) holds significant economic importance globally. Evaluating a diverse range of germplasm based on molecular characteristics not only enhances its preservation but also supports its utilization in breeding programs. In this study, we assessed genetic diversity and population structure among 155 cassava genotypes from Uganda using 5247 single nucleotide polymorphism (SNP) markers. Genotyping by sequencing (GBS) was employed for SNP discovery and to evaluate genetic diversity and population structure using the ADMIXTURE software. The cassava accessions comprised two populations: 49 accessions from Ugandan lines and 106 accessions resulting from crosses between South American and Ugandan lines. The average call rate of 96% was utilized to assess marker polymorphism. Polymorphic information content values of the markers ranged from 0.1 to 0.5 with an average of 0.4 which was moderately high. The principal component analysis (PCA) showed that the first two components captured ~ 24.2% of the genetic variation. The average genetic diversity was 0.3. The analysis of molecular variance (AMOVA) indicated that 66.02% and 33.98% of the total genetic variation occurred within accessions and between sub-populations, respectively. Five sub-populations were identified based on ADMIXTURE structure analysis (K = 5). Neighbor-joining tree and hierarchical clustering tree revealed the presence of three different groups which were primarily based on the source of the genotypes. The results suggested that there was considerable genetic variation among the cassava genotypes which is useful in cassava improvement and conservation efforts.

Exploring Burnout among Nursing Students in Bangalore: A t-Distributed Stochastic Neighbor Embedding Analysis and Hierarchical Clustering in Cross-Sectional Data.

This study explores burnout among nursing students in Bangalore, India, focusing on Exhaustion and Disengagement scores. A cross-sectional design was applied using the Oldenburg Burnout Inventory modified for nursing students, collecting data using a survey that was conducted between October and December 2023. The sample consisted of 237 female nursing students from the Bachelor of Science in Nursing program at Bangalore College of Nursing, South India. The study integrated the t-distributed Stochastic Neighbor Embedding (t-SNE) procedure for data simplification into three t-SNE components, used in a hierarchical clustering analysis, which identified distinct student profiles: "High-Intensity Study Group" and "Altruistic Aspirants". While burnout scores were generally high, students with high study hours ("High-Intensity Study Group") reported greater Exhaustion, with a mean score of 26.78 (SD = 5.26), compared to those in the "Altruistic Aspirants" group, who reported a mean score of 25.00 (SD = 4.48), demonstrating significant differences (p-value = 0.005). Conversely, those motivated by altruism ("Altruistic Aspirants") showed higher Disengagement, with a mean score of 19.78 (SD = 5.08), in contrast to "High-Intensity Study Group", which reported a lower mean of 17.84 (SD = 4.74) (p-value = 0.002). This segmentation suggests that burnout manifests differently depending on the students' academic load and intrinsic motivations. This study underscores the need for targeted interventions that address specific factors characterizing the clusters and provide information for designing future research and interventions. This study was not registered.

[Cluster Analysis and Ablation Success Rate in Atrial Fibrillation Patients Undergoing Catheter Ablation]. / æˆ¿é¢¤å°„é¢‘æ¶ˆèžæ‚£è€ çš„èšç±»åˆ†æžåŠæ¶ˆèžæˆåŠŸçŽ‡è¯„ä»·.

Objective: Atrial fibrillation (AF) is a disease of high heterogeneity, and the association between AF phenotypes and the outcome of different catheter ablation strategies remains unclear. Conventional classification of AF (e.g. according to duration, atrial size, and thromboembolism risk) fails to provide reference for the optimal stratification of the prognostic risks or to guide individualized treatment plan. In recent years, research on machine learning has found that cluster analysis, an unsupervised data-driven approach, can uncover the intrinsic structure of data and identify clusters of patients with pathophysiological similarity. It has been demonstrated that cluster analysis helps improve the characterization of AF phenotypes and provide valuable prognostic information. In our cohort of AF inpatients undergoing radiofrequency catheter ablation, we used unsupervised cluster analysis to identify patient subgroups, to compare them with previous studies, and to evaluate their association with different suitable ablation patterns and outcomes. Methods: The participants were AF patients undergoing radiofrequency catheter ablation at West China Hospital between October 2015 and December 2017. All participants were aged 18 years or older. They underwent radiofrequency catheter ablation during their hospitalization. They completed the follow-up process under explicit informed consent. Patients with AF of a reversible cause, severe mitral stenosis or prosthetic heart valve, congenital heart disease, new-onset acute coronary syndrome within three months prior to the surgery, or a life expectancy less than 12 months were excluded according to the exclusion criteria. The cohort consisted of 1102 participants with paroxysmal or persistent/long-standing persistent AF. Data on 59 variables representing demographics, AF type, comorbidities, therapeutic history, vital signs, electrocardiographic and echocardiographic findings, and laboratory findings were collected. Overall, data for the variables were rarely missing (<5%), and multiple imputation was used for correction of missing data. Follow-up surveys were conducted through outpatient clinic visits or by telephone. Patients were scheduled for follow-up with 12-lead resting electrocardiography and 24-hours Holter monitoring at 3 months and 6 months after the ablation procedure. Early ablation success was defined as the absence of documented AF, atrial flutter, or atrial tachycardia >30 seconds at 6-month follow-up. Hierarchical clustering was performed on the 59 baseline variables. All characteristic variables were standardized to have a mean of zero and a standard deviation of one. Initially, each patient was regarded as a separate cluster, and the distance between these clusters was calculated. Then, the Ward minimum variance method of clustering was used to merge the pair of clusters with the minimum total variance. This process continued until all patients formed one whole cluster. The "NbClust" package in R software, capable of calculating various statistical indices, including pseudo t2 index, cubic clustering criterion, silhouette index etc, was applied to determine the optimal number of clusters. The most frequently chosen number of clusters by these indices was selected. A heatmap was generated to illustrate the clinical features of clusters, while a tree diagram was used to depict the clustering process and the heterogeneity among clusters. Ablation strategies were compared within each cluster regarding ablation efficacy. Results: Five statistically driven clusters were identified: 1) the younger age cluster (n=404), characterized by the lowest prevalence of cardiovascular and cerebrovascular comorbidities but the highest prevalence of obstructive sleep apnea syndrome (14.4%); 2) a cluster of elderly adults with chronic diseases (n=438), the largest cluster, showing relatively higher rates of hypertension, diabetes, stroke, and chronic obstructive pulmonary disease; 3) a cluster with high prevalence of sinus node dysfunction (n=160), with patients showing the highest prevalence of sick sinus syndrome and pacemaker implantation; 4) the heart failure cluster (n=80), with the highest prevalence of heart failure (58.8%) and persistent/long-standing persistent AF (73.7%); 5) prior coronary artery revascularization cluster (n=20), with patients of the most advanced age (median: 69.0 years old) and predominantly male patients, all of whom had prior myocardial infarction and coronary artery revascularization. Patients in cluster 2 achieved higher early ablation success with pulmonary veins isolation alone compared to extensive ablation strategies (79.6% vs. 66.5%; odds ratio [OR]=1.97, 95% confidence interval [CI]: 1.28-3.03). Although extensive ablation strategies had a slightly higher success rate in the heart failure group, the difference was not statistically significant. Conclusions: This study provided a unique classification of AF patients undergoing catheter ablation by cluster analysis. Age, chronic disease, sinus node dysfunction, heart failure and history of coronary artery revascularization contributed to the formation of the five clinically relevant subtypes. These subtypes showed differences in ablation success rates, highlighting the potential of cluster analysis in guiding individualized risk stratification and treatment decisions for AF patients.

Geotemporal Analysis of Melanoma Incidence and Mortality Identifies Distinct Clusters of Trends within the United States Between 2001 and 2019.

The extent to which the geographic diversity of the US plays a significant role in melanoma incidence and mortality over time has not been precisely characterized. We obtained age-adjusted melanoma data for the 50 states between 2001 and 2019 from the Surveillance, Epidemiology, and End Results registry and performed hierarchical clustering (complete linkage, Euclidean space) to uncover geo-temporal trend groups over 2 decades. While there was a global increase in incidence during this time (b1 = +0.41, P < .0001), there were 6 distinct clusters (by absolute and Z-score) with significantly different temporal trends (analysis of covariance P < .0001). Cluster 2 states had the sharpest increase in incidence with b1 = +0.66, P < .0001. For mortality, the global rate decreased (b1 = -0.03, P = .0003) with 3 and 6 clusters by absolute and Z-scores, respectively (analysis of covariance P < .05). Cluster 1 states exhibited the smallest decline in mortality (b1 = -0.017, P = .008). Mortality to incidence ratios declined (b1 = -0.0037, P < .0001) and harbored 4 and 6 clusters by absolute and Z-score analysis, respectively (analysis of covariance P < .0001). Cluster 4 states had the lowest rate of mortality to incidence ratios decline (b1 = -0.003, P < .0001). These results provide an unprecedented higher dimensional view of melanoma behavior over space and time. With more refined analyses, geospatial studies can uncover local trends which can inform public health agencies to more properly allocate resources.

The cluster D-trace loss for differential network analysis.

A growing literature suggests that gene expression can be greatly altered in disease conditions, and identifying those changes will improve the understanding of complex diseases such as cancers or diabetes. A prevailing direction in the analysis of gene expression studies the changes in gene pathways which include sets of related genes. Therefore, introducing structured exploration to differential analysis of gene expression networks may lead to meaningful discoveries. The topic of this paper is differential network analysis, which focuses on capturing the differences between two or more precision matrices. We discuss the connection between the thresholding method and the D-trace loss method on differential network analysis in the case that the precision matrices share the common connected components. Based on this connection, we further propose the cluster D-trace loss method which directly estimates the differential network and achieves model selection consistency. Simulation studies demonstrate its improved performance and computational efficiency. Finally, the usefulness of our proposed estimator is demonstrated by a real-data analysis on non-small cell lung cancer.

Multivariate Analysis of Solubility Parameters for Drug-Polymer Miscibility Assessment in Preparing Raloxifene Hydrochloride Amorphous Solid Dispersions.

The success of obtaining solid dispersions for solubility improvement invariably depends on the miscibility of the drug and polymeric carriers. This study aimed to categorize and select polymeric carriers via the classical group contribution method using the multivariate analysis of the calculated solubility parameter of RX-HCl. The total, partial, and derivate parameters for RX-HCl were calculated. The data were compared with the results of excipients (N = 36), and a hierarchical clustering analysis was further performed. Solid dispersions of selected polymers in different drug loads were produced using solvent casting and characterized via X-ray diffraction, infrared spectroscopy and scanning electron microscopy. RX-HCl presented a Hansen solubility parameter (HSP) of 23.52 MPa1/2. The exploratory analysis of HSP and relative energy difference (RED) elicited a classification for miscible (n = 11), partially miscible (n = 15), and immiscible (n = 10) combinations. The experimental validation followed by a principal component regression exhibited a significant correlation between the crystallinity reduction and calculated parameters, whereas the spectroscopic evaluation highlighted the hydrogen-bonding contribution towards amorphization. The systematic approach presented a high discrimination ability, contributing to optimal excipient selection for the obtention of solid solutions of RX-HCl.

Phenotypes of Polish primary care patients using hierarchical clustering: Exploring the risk of mortality in the LIPIDOGEN2015 study cohort.

BACKGROUND: Comorbidities in primary care do not occur in isolation but tend to cluster together causing various clinically complex phenotypes. This study aimed to distinguish phenotype clusters and identify the risks of all-cause mortality in primary care. METHODS: The baseline cohort of the LIPIDOGEN2015 sub-study involved 1779 patients recruited by 438 primary care physicians. To identify different phenotype clusters, we used hierarchical clustering and investigated differences between clinical characteristics and mortality between clusters. We then performed causal analyses using causal mediation analysis to explore potential mediators between different clusters and all-cause mortality. RESULTS: A total of 1756 patients were included (mean age 51.2, SD 13.0; 60.3% female), with a median follow-up of 5.7 years. Three clusters were identified: Cluster 1 (n = 543) was characterised by overweight/obesity (body mass index ≥ 25 kg/m2), older (age ≥ 65 years), more comorbidities; Cluster 2 (n = 459) was characterised by non-overweight/obesity, younger, fewer comorbidities; Cluster 3 (n = 754) was characterised by overweight/obesity, younger, fewer comorbidities. Adjusted Cox regression showed that compared with Cluster 2, Cluster 1 had a significantly higher risk of all-cause mortality (HR 3.87, 95% CI: 1.24-15.91), whereas this was insignificantly different for Cluster 3. Causal mediation analyses showed that decreased protein thiol groups mediated the hazard effect of all-cause mortality in Cluster 1 compared with Cluster 2, but not between Clusters 1 and 3. CONCLUSION: Overweight/obesity older patients with more comorbidities had the highest risk of long-term all-cause mortality, and in the young group population overweight/obesity insignificantly increased the risk in the long-term follow-up, providing a basis for stratified phenotypic risk management.

Characterisation of colorectal cancer by hierarchical clustering analyses for five stroma-related markers.

Evidence for the tumour-supporting capacities of the tumour stroma has accumulated rapidly in colorectal cancer (CRC). Tumour stroma is composed of heterogeneous cells and components including cancer-associated fibroblasts (CAFs), small vessels, immune cells, and extracellular matrix proteins. The present study examined the characteristics of CAFs and collagen, major components of cancer stroma, by immunohistochemistry and Sirius red staining. The expression status of five independent CAF-related or stromal markers, decorin (DCN), fibroblast activation protein (FAP), podoplanin (PDPN), alpha-smooth muscle actin (ACTA2), and collagen, and their association with clinicopathological features and clinical outcomes were analysed. Patients with DCN-high tumours had a significantly worse 5-year survival rate (57.3% versus 79.0%; p = 0.044). Furthermore, hierarchical clustering analyses for these five markers identified three groups that showed specific characteristics: a solid group (cancer cell-rich, DCNLowPDPNLow); a PDPN-dominant group (DCNMidPDPNHigh); and a DCN-dominant group (DCNHighPDPNLow), with a significant association with patient survival (p = 0.0085). Cox proportional hazards model identified the PDPN-dominant group (hazard ratio = 0.50, 95% CI = 0.26-0.96, p = 0.037) as a potential favourable factor compared with the DCN-dominant group. Of note, DCN-dominant tumours showed the most advanced pT stage and contained the lowest number of CD8+ and FOXP3+ immune cells. This study has revealed that immunohistochemistry and special staining of five stromal factors with hierarchical clustering analyses could be used for the prognostication of patients with CRC. Cancer stroma-targeting therapies may be candidate treatments for patients with CRC.

Clustering plasma concentration-time curves: applications of unsupervised learning in pharmacogenomics.

Pharmaceutical researchers are continually searching for techniques to improve both drug development processes and patient outcomes. An area of recent interest is the potential for machine learning (ML) applications within pharmacology. One such application not yet given close study is the unsupervised clustering of plasma concentration-time curves, hereafter, pharmacokinetic (PK) curves. In this paper, we present our findings on how to cluster PK curves by their similarity. Specifically, we find clustering to be effective at identifying similar-shaped PK curves and informative for understanding patterns within each cluster of PK curves. Because PK curves are time series data objects, our approach utilizes the extensive body of research related to the clustering of time series data as a starting point. As such, we examine many dissimilarity measures between time series data objects to find those most suitable for PK curves. We identify Euclidean distance as generally most appropriate for clustering PK curves, and we further show that dynamic time warping, Fréchet, and structure-based measures of dissimilarity like correlation may produce unexpected results. As an illustration, we apply these methods in a case study with 250 PK curves used in a previous pharmacogenomic study. Our case study finds that an unsupervised ML clustering with Euclidean distance, without any subject genetic information, is able to independently validate the same conclusions as the reference pharmacogenomic results. To our knowledge, this is the first such demonstration. Further, the case study demonstrates how the clustering of PK curves may generate insights that could be difficult to perceive solely with population level summary statistics of PK metrics.

Whole-Genome Sequencing of Human and Porcine Escherichia coli Isolates on a Commercial Pig Farm in South Africa.

Escherichia coli is an indicator micro-organism in One Health antibiotic resistance surveillance programs. The purpose of the study was to describe and compare E. coli isolates obtained from pigs and human contacts from a commercial farm in South Africa using conventional methods and whole-genome sequencing (WGS). Porcine E. coli isolates were proportionally more resistant phenotypically and harbored a richer diversity of antibiotic resistance genes as compared to human E. coli isolates. Different pathovars, namely ExPEC (12.43%, 21/169), ETEC (4.14%, 7/169), EPEC (2.96%, 5/169), EAEC (2.96%, 5/169) and STEC (1.18%, 2/169), were detected at low frequencies. Sequence type complex (STc) 10 was the most prevalent (85.51%, 59/169) among human and porcine isolates. Six STcs (STc10, STc86, STc168, STc206, STc278 and STc469) were shared at the human-livestock interface according to multilocus sequence typing (MLST). Core-genome MLST and hierarchical clustering (HC) showed that human and porcine isolates were overall genetically diverse, but some clustering at HC2-HC200 was observed. In conclusion, even though the isolates shared a spatiotemporal relationship, there were still differences in the virulence potential, antibiotic resistance profiles and cgMLST and HC according to the source of isolation.

[Geographical division of terrestrial mammals in China based on species composition characteristics]. / åŸºäºŽç‰©ç§æž„æˆç‰¹å¾çš„ä¸­å›½é™†æ –å“ºä¹³åŠ¨ç‰©åœ°ç†åˆ†åŒº.

China has complex natural conditions and is rich in biodiversity. Based on the geographical distribution and species composition of terrestrial mammals, we explored the characteristics and geographic partitioning of mammal populations in different regions of China. We used a clustering algorithm, combined with the spatial distribution data and taxonomic characteristics of mammals, to geographically partition the terrestrial mammals in China. We found 10 zoogeographic regions of terrestrial mammals in China: Northeast region, North China region, Eastern grassland region, Western region, Northwest region, Qiangtang plateau region, Eastern Qinghai-Tibet Plateau region, Himalayan region, South China region, and Taiwan-Hainan region. We found a new geographical zoning pattern for terrestrial mammals in China, examined the variability and characteristics of species composition among different regions, and quantified the association between species distribution and environmental factors. We proposed a method of incorporating taxonomic information into cluster analysis, which provided a new idea for zoogeographic region studies, a new perspective for understanding species diversity, and a scientific basis for animal conservation and habitat planning.

KEMoS: A knowledge-enhanced multi-modal summarizing framework for Chinese online meetings.

The demand for "online meetings" and "collaborative office work" keeps surging recently, producing an abundant amount of relevant data. How to provide participants with accurate and fast summarizing service has attracted extensive attention. Existing meeting summarizing models overlook the utilization of multi-modal information and the information offsetting during summarizing. In this paper, we develop a knowledge-enhanced multi-modal summarizing framework. Firstly, we construct a three-layer multi-modal meeting knowledge graph, including basic, knowledge, and multi-modal layer, to integrate meeting information thoroughly. Then, we raise a topic-based hierarchical clustering approach, which considers information entropy and difference simultaneously, to capture the semantic evolution of meetings. Next, we devise a multi-modal enhanced encoding strategy, including a sentence-level cross-modal encoder, a joint loss function, and a knowledge graph embedding module, to learn the meeting and topic-level presentations. Finally, when generating summaries, we design a topic-enhanced decoding strategy for the Transformer decoder which mitigates semantic offsetting with the aid of topic information. Extensive experiments show that our proposed work consistently outperforms state-of-the-art solutions on the Chinese meeting dataset, where the ROUGE-1, ROUGE-2, and ROUGE-L are 49.98%, 21.03%, and 32.03% respectively.

Role of prior immunity in binding to spike of "future" Omicron subvariants.

BACKGROUND: Throughout the COVID-19 pandemic, virus evolution and large-scale vaccination programs have caused multiple exposures to SARS CoV-2 spike protein, resulting in complex antibody profiles. The binding of these to spike protein of "future" variants in the context of such heterogeneous exposure has not been studied. METHODS: We tested archival sera (Delta and Omicron period) stratified by anti-spike antibody (including IgG) levels for reactivity to Omicron-subvariants(BA.1, BA.2,BA.2.12.1, BA.2.75, BA.4/5 and BF.7) spike protein. Assessed antigenic distance between groups using Antigenic Cartography and performed hierarchical clustering of antibody data in a Euclidean distance framework. RESULTS: Antibody (including IgG) antibody reactivity to Wild-type (CLIA) and subvariants (ELISA) spike protein were similar between periods (p > 0.05). Both 'High S' and 'Low S' of Delta and Omicron periods were closely related to "future" subvariants by Antigenic Cartography. Sera from different S groups clustered together with 'Low S' interspersed between 'High S' on hierarchical clustering, suggesting common binding sites. Further, anti-spike antibodies (including IgG) to Wild-type (S1/S2 and Trimeric S) clustered with Omicron-subvariant binding antibodies. CONCLUSIONS: Hybrid immunity caused by cumulative virus exposure in Delta or Omicron periods resulted in equivalent binding to "future" variants, which might be due to binding to conserved regions of spike protein of future variants. A prominent finding is that the 'Low S' antibody demonstrates similar binding.

LcDel: deletion variation detection based on clustering and long reads.

Motivation: Genomic structural variation refers to chromosomal level variations such as genome rearrangement or insertion/deletion, which typically involve larger DNA fragments compared to single nucleotide variations. Deletion is a common type of structural variants in the genome, which may lead to mangy diseases, so the detection of deletions can help to gain insights into the pathogenesis of diseases and provide accurate information for disease diagnosis, treatment, and prevention. Many tools exist for deletion variant detection, but they are still inadequate in some aspects, and most of them ignore the presence of chimeric variants in clustering, resulting in less precise clustering results. Results: In this paper, we present LcDel, which can detect deletion variation based on clustering and long reads. LcDel first finds the candidate deletion sites and then performs the first clustering step using two clustering methods (sliding window-based and coverage-based, respectively) based on the length of the deletion. After that, LcDel immediately uses the second clustering by hierarchical clustering to determine the location and length of the deletion. LcDel is benchmarked against some other structural variation detection tools on multiple datasets, and the results show that LcDel has better detection performance for deletion. The source code is available in https://github.com/cyq1314woaini/LcDel.

Visualization strategies to aid interpretation of high-dimensional genotoxicity data.

This article describes a range of high-dimensional data visualization strategies that we have explored for their ability to complement machine learning algorithm predictions derived from MultiFlow® assay results. For this exercise, we focused on seven biomarker responses resulting from the exposure of TK6 cells to each of 126 diverse chemicals over a range of concentrations. Obviously, challenges associated with visualizing seven biomarker responses were further complicated whenever there was a desire to represent the entire 126 chemical data set as opposed to results from a single chemical. Scatter plots, spider plots, parallel coordinate plots, hierarchical clustering, principal component analysis, toxicological prioritization index, multidimensional scaling, t-distributed stochastic neighbor embedding, and uniform manifold approximation and projection are each considered in turn. Our report provides a comparative analysis of these techniques. In an era where multiplexed assays and machine learning algorithms are becoming the norm, stakeholders should find some of these visualization strategies useful for efficiently and effectively interpreting their high-dimensional data.

Safety in higher level automated vehicles: Investigating edge cases in crashes of vehicles equipped with automated driving systems.

With emerging Automated Driving Systems (ADS) representing Automated Vehicles (AVs) of Level 3 or higher as classified by the Society of Automotive Engineers, several AV manufacturers are testing their vehicles on public roadways in the U.S. The safety performance of AVs has become a major concern for the transportation industry. Several ADS-equipped vehicle crashes have been reported to the National Highway Traffic Safety Administration (NHTSA) in recent years. Scrutinizing these crashes can reveal rare or complex scenarios beyond the normal capabilities of AV technologies called "edge cases." Investigating edge-case crashes helps AV companies prepare vehicles to handle these unusual scenarios and, as such, improves traffic safety. Through analyzing the NHTSA data from July 2021 to February 2023, this study utilizes an unsupervised machine learning technique, hierarchical clustering, to identify edge cases in ADS-equipped vehicle crashes. Fifteen out of 189 observations are identified as edge cases, representing 8 % of the population. Injuries occurred in 10 % of all crashes (19 out of 189), but the proportion rose to 27 % for edge cases (4 out of 15 edge cases). Based on the results, edge cases could be initiated by AVs, humans, infrastructure/environment, or their combination. Humans can be identified as one of the contributors to the onset of edge-case crashes in 60 % of the edge cases (9 out of 15 edge cases). The main scenarios for edge cases include unlawful behaviors of crash partners, absence of a safety driver within the AV, precrash disengagement, and complex events challenging for ADS, e.g., unexpected obstacles, unclear road markings, and sudden and unexpected changes in traffic flow, such as abrupt road congestion or sudden stopped traffic from a crash. Identifying and investigating edge cases is crucial for improving transportation safety and building public trust in AVs.

Brain topology improved spiking neural network for efficient reinforcement learning of continuous control.

The brain topology highly reflects the complex cognitive functions of the biological brain after million-years of evolution. Learning from these biological topologies is a smarter and easier way to achieve brain-like intelligence with features of efficiency, robustness, and flexibility. Here we proposed a brain topology-improved spiking neural network (BT-SNN) for efficient reinforcement learning. First, hundreds of biological topologies are generated and selected as subsets of the Allen mouse brain topology with the help of the Tanimoto hierarchical clustering algorithm, which has been widely used in analyzing key features of the brain connectome. Second, a few biological constraints are used to filter out three key topology candidates, including but not limited to the proportion of node functions (e.g., sensation, memory, and motor types) and network sparsity. Third, the network topology is integrated with the hybrid numerical solver-improved leaky-integrated and fire neurons. Fourth, the algorithm is then tuned with an evolutionary algorithm named adaptive random search instead of backpropagation to guide synaptic modifications without affecting raw key features of the topology. Fifth, under the test of four animal-survival-like RL tasks (i.e., dynamic controlling in Mujoco), the BT-SNN can achieve higher scores than not only counterpart SNN using random topology but also some classical ANNs (i.e., long-short-term memory and multi-layer perception). This result indicates that the research effort of incorporating biological topology and evolutionary learning rules has much in store for the future.

Tracking and Profiling Repeated Users Over Time in Text-Based Counseling: Longitudinal Observational Study With Hierarchical Clustering.

BACKGROUND: Due to their accessibility and anonymity, web-based counseling services are expanding at an unprecedented rate. One of the most prominent challenges such services face is repeated users, who represent a small fraction of total users but consume significant resources by continually returning to the system and reiterating the same narrative and issues. A deeper understanding of repeated users and tailoring interventions may help improve service efficiency and effectiveness. Previous studies on repeated users were mainly on telephone counseling, and the classification of repeated users tended to be arbitrary and failed to capture the heterogeneity in this group of users. OBJECTIVE: In this study, we aimed to develop a systematic method to profile repeated users and to understand what drives their use of the service. By doing so, we aimed to provide insight and practical implications that can inform the provision of service catering to different types of users and improve service effectiveness. METHODS: We extracted session data from 29,400 users from a free 24/7 web-based counseling service from 2018 to 2021. To systematically investigate the heterogeneity of repeated users, hierarchical clustering was used to classify the users based on 3 indicators of service use behaviors, including the duration of their user journey, use frequency, and intensity. We then compared the psychological profile of the identified subgroups including their suicide risks and primary concerns to gain insights into the factors driving their patterns of service use. RESULTS: Three clusters of repeated users with clear psychological profiles were detected: episodic, intermittent, and persistent-intensive users. Generally, compared with one-time users, repeated users showed higher suicide risks and more complicated backgrounds, including more severe presenting issues such as suicide or self-harm, bullying, and addictive behaviors. Higher frequency and intensity of service use were also associated with elevated suicide risk levels and a higher proportion of users citing mental disorders as their primary concerns. CONCLUSIONS: This study presents a systematic method of identifying and classifying repeated users in web-based counseling services. The proposed bottom-up clustering method identified 3 subgroups of repeated users with distinct service behaviors and psychological profiles. The findings can facilitate frontline personnel in delivering more efficient interventions and the proposed method can also be meaningful to a wider range of services in improving service provision, resource allocation, and service effectiveness.