Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “unsupervised clustering”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 307 records · Page 17Linked to original sources
PubMed2026

A per- and polyfluoroalkyl substances-based gene signature links prognosis to immune landscapes in thyroid cancer.

BACKGROUND: Thyroid cancer (THCA) is the most common endocrine malignancy with a rising global incidence and significant heterogeneity. Although per- and polyfluoroalkyl substances (PFAS) exposure is linked to thyroid dysfunction, the prognostic value of per- and polyfluoroalkyl substances-related genes (PFASRGs) and their role in the tumor immune microenvironment (TME) remain poorly understood. This study aims to systematically screen key PFASRGs and evaluate their prognostic value as biomarkers for THCA. METHODS: Utilizing The Cancer Genome Atlas (TCGA)-THCA transcriptomic data and PFASRGs, we constructed a prognostic model through differential expression analysis, univariate and multivariate Cox regression analyses, and the least absolute shrinkage and selection operator (LASSO). The model's robustness was validated using receiver operating characteristic (ROC) curves, Kaplan-Meier analysis, and clinical nomograms. Furthermore, the TME, immunotherapy response, and drug sensitivities were systematically evaluated. Distinct molecular landscapes were characterized by stratifying the cohort via unsupervised consensus clustering analysis. RESULTS: The eight-gene prognostic model demonstrated robust performance, with area under the curve (AUC) values exceeding 0.85 across all validation cohorts. High-risk patients exhibited significantly shorter overall survival and an "inflamed" TME characterized by high immune scores and checkpoint expression. In contrast, the therapeutic efficacy of anti-programmed death-ligand 1 (PD-L1) agents was more pronounced in the low-risk category, as evidenced by a superior objective response. Furthermore, distinct molecular subtypes and risk-specific sensitivities to targeted agents, such as sorafenib and sunitinib, were identified, highlighting the model's clinical utility for personalized treatment. CONCLUSIONS: We established a novel THCA prognostic framework based on eight PFASRGs. This model exhibits superior performance in risk stratification, effectively distinguishing cohorts with divergent clinical trajectories, unique immune microenvironment features, and varied therapeutic responses. Our findings provide a powerful predictive tool for refining prognostic evaluation and facilitating the implementation of personalized management strategies for THCA patients.

Per- and polyfluoroalkyl substances-related genes ↗
PubMed2006

Microarray analysis of thyroid nodule fine-needle aspirates accurately classifies benign and malignant lesions.

Current preoperative diagnostic procedures for thyroid nodules rely mainly on the cytological interpretation of fine-needle aspirates (FNAs). DNA microarray analysis has been shown to reliably distinguish benign and malignant thyroid nodules in surgically resected specimens, but its diagnostic potential in thyroid FNA has not been examined. In the present study, the expression profiles of 50 benign thyroid lesions and papillary thyroid carcinoma tissue samples were compared, generating a list of 25 differentially expressed genes from this training set. A test set of 22 FNA specimens was evaluated by unsupervised hierarchical cluster analysis using this gene list, and the results were compared to FNA cytology. FNA specimens were found to fall into three clusters: malignant (n = 10), benign (n = 7), and indeterminate (n = 5). The benign and malignant groups showed complete concordance with the final histological diagnosis except for one histologically benign lesion, which was rediagnosed as follicular variant of papillary thyroid carcinoma on histological review. Paired analysis between FNA and matched tissues samples illustrated adequate sampling with FNA. These results illustrate that microarray analysis of FNA is feasible and has the potential to improve the accuracy of FNA in categorizing benign from malignant lesions beyond routine cytological evaluation.

Biopsy, Fine-Needle↗
PubMed2026

Integrative subtyping by bile acid metabolism identifies CLCA1/UGT2A3/ZG16 as markers of immune dysfunction and poor prognosis in colorectal cancer.

BACKGROUND: Colorectal cancer (CRC) is the primary driver of cancer-related death and illness across the world. Despite the full-scale shift of the treatment approach for some colorectal cancer patients due to the use of immune checkpoint inhibitors (ICIs), primary resistance still poses a huge challenge to clinicians. Bile acid metabolism is involved in the pathogenesis of CRC. However, its particular function in shaping the tumor immune microenvironment (TIME) and its effect on prognosis and immune treatment response remain unclear. METHODS: Based on the transcriptome and clinical data from The Cancer Genome Atlas-Colon Adenocarcinoma (TCGA-COAD) cohort, we performed unsupervised consensus clustering and classified patients into different molecular subtypes according to bile acid metabolism. We subsequently compared overall survival (OS), immune cell infiltration levels, and differentially expressed genes among the subtypes. In addition, protein-protein interaction (PPI) network and Cox proportional hazards regression were used to identify key hub genes. Finally, the expression of these crucial hub genes was validated in the Gene Expression Omnibus (GEO) cohort and independent clinical patients. RESULTS: The bile-low group showed a significant reduction in OS time (p = 0.0049). The infiltration levels of CD8+ T cells (p < 0.05) and M1 macrophages (p < 0.01) were significantly higher in the bile-low group than in the bile-high group. We identified three key genes-CLCA1, UGT2A3, and ZG16-and found that they all were downregulated in tumor tissues across the TCGA-COAD and GEO datasets, as well as in independent clinical samples. Survival analysis showed that high CLCA1 expression was significantly associated with favorable overall survival (p < 0.001), whereas UGT2A3 (p = 0.23) and ZG16 (p = 0.17) did not reach statistical significance. The three hub genes were negatively correlated with the (TIDE) score (CLCA1: R = - 0.24, p < 0.001; UGT2A3: R = - 0.15, p = 0.0022; ZG16: R = - 0.14, p = 0.0039). CONCLUSION: Our findings suggest that bile acid metabolism could shape the TIME via key genes CLCA1, UGT2A3, and ZG16, and subsequently modify CRC prognosis and immunotherapy responses. These genes may serve as potential prognostic indicators and mechanistic mediators linking bile acid metabolism to T-cell dysfunction, offering insights for future combination strategies targeting the metabolism-barrier-immunity axis.

CLCA1↗
PubMed2003

Tumor classification based on gene expression profiling shows that uveal melanomas with and without monosomy 3 represent two distinct entities.

Uveal melanoma is the most common intraocular malignancy. About 50% of patients die of metastases, which almost exclusively originate from primary tumors that have lost one chromosome 3 (monosomy 3). To gain insight into the biological mechanisms that underlie the various metastasizing potential of uveal melanoma, we have determined gene expression levels in 20 primary tumors using oligonucleotide microarrays containing 12500 probe sets. The expression measurements of those 7902 genes that were expressed in more than 10% of tumors were analyzed using two different statistical approaches. We used a modified Wilcoxon rank-sum test to identify genes differentially expressed between tumors with and without monosomy 3. Seven genes showed complete loss of expression in tumors with monosomy 3 but were expressed in tumors with disomy 3. Two of them, CHL1 and fls485, are located within or close to the uveal melanoma susceptibility locus UVM2 at 3p25. However, mutation analysis of both genes in eight tumors with monosomy 3 did not reveal structural or epigenetic alteration. To identify tumor classes, we performed unsupervised hierarchical cluster analysis; this approach separated uveal melanomas into two groups. We found that this classification is strikingly robust because, when tested by "resampling," the same grouping is obtained from 47 of 50 subsamples of genes. In clusterings of the three remaining subsamples, the grouping of only one tumor does not conform with the original classification. Excluding this tumor, cluster analyses of subsamples containing as few as 300 randomly chosen genes consistently result in the same classification, thus indicating that the difference between the two tumor classes is pervasive. Interestingly, all of the tumors in one of the groups have disomy 3, whereas all of the others have monosomy 3. Our findings suggest that there are two distinct entities of uveal melanoma that were previously unrecognized because they are not obviously distinguishable by clinicopathological features.

Chromosomes, Human, Pair 3↗
PubMed2004

Genome-wide gene-expression profiles of breast-cancer cells purified with laser microbeam microdissection: identification of genes associated with progression and metastasis.

Breast carcinoma is a complex disease characterized by accumulation of multiple genetic alterations, and the understanding of the molecular basis of mammary tumorigenesis is still incomplete. In this study we analyzed gene-expression profiles of 81 surgical specimens of 12 ductal carcinoma in situ (DCIS) and 69 invasive ductal carcinoma (IDC). After applying laser-microbeam micro-dissection to all samples we achieved 98-99% pure populations of breast cancer cells, and of normal breast epithelial cells used as controls. A cDNA-microarray analysis of 23,040 genes in these samples and a subsequent unsupervised hierarchical clustering distinguished two tumor groups, mainly in terms of estrogen-receptor (ER) status. We then undertook a supervised analysis and identified 325 genes that were commonly either up- or down-regulated in both pathologically discrete stages (DCIS and IDC), indicating that these genes might play important roles in malignant transformation of breast ductal cells. In addition, we searched invasion-associated gene candidates whose expression was altered in IDC, but not in DCIS, and identified 24 up-regulated genes and 41 down-regulated genes. Furthermore, we identified 34 genes that were expressed differently in tumors from patients with lymph node metastasis as opposed to no metastasis. On that basis we developed a scoring system that correlated well with the metastatic status. Tumors from all of the 37 test patients with lymph-node metastasis yielded positive scores by our definition, whereas 38 of the 40 tumors (95%) without lymph node metastasis had negative scores. Our data should provide useful information for identifying predictive markers for invasion or metastasis, and suggest potential target molecules for treatment of breast cancers.

Adult↗
PubMed2004

Combined genetic and transcriptional profiling of acute myeloid leukemia with normal and complex karyotypes.

BACKGROUND AND OBJECTIVES: Acute myeloid leukemia (AML) is a heterogeneous group of diseases. Patients with a normal karyotype constitute the largest single group; multiple chromosome rearrangements involving three or more chromosomes occur in 5 10% of AML patients. The pathophysiologic mechanisms underlying both groups are largely unknown. In the current study, we have systematically combined transcriptional profiles with cytogenetic data from 15 AML patients with either normal or complex karyotypes. DESIGN AND METHODS: The expression profiles were investigated by unsupervised hierarchical clustering, supervised cluster analysis, and comparative genomic microarray analysis. In addition, the samples were analyzed by G-banding and/or spectral karyotyping and comparative genomic hybridization. RESULTS: Our results show that AML with complex karyotypes exhibit a gene expression profile that is specific to this group of patients. The differentially expressed genes included several located on 5q and 7q, as well as genes involved in controlling cell division. We also found that DNA gains and losses caused by multiple chromosome rearrangements result in altered gene expression in a gene-dosage-dependent manner. INTERPRETATION AND CONCLUSIONS: These data provide insight into the mechanisms of multiple chromosome rearrangements and further demonstrate that the expression patterns of AML are strongly linked to the karyotypic status, even for the relatively undefined cytogenetic subgroup AML with complex karyotype.

Acute Disease↗
PubMed2005

Transcriptional profiles in peripheral blood mononuclear cells prognostic of clinical outcomes in patients with advanced renal cell carcinoma.

PURPOSE: Given their accessibility, surrogate tissues, such as peripheral blood mononuclear cells (PBMC), may provide potential predictive biomarkers in clinical pharmacogenomic studies. In leukemias and lymphomas, the prognostic value of peripheral blast expression profiles is clear; however, it is unclear whether circulating mononuclear cells of patients with solid tumors might yield profiles with similar prognostic associations. EXPERIMENTAL DESIGN: In this study, we evaluated the association of expression profiles in PBMCs with clinical outcomes in patients with advanced renal cell cancer. Transcriptional patterns in PBMCs of 45 renal cell cancer patients were compared with clinical outcome data at the conclusion of a phase II study of the mTOR kinase inhibitor CCI-779 to determine whether pretreatment transcriptional patterns in PBMCs were correlated with eventual patient outcomes. RESULTS: Unsupervised hierarchical clustering of the PBMC profiles using all expressed genes identified clusters of patients with significant differences in survival. Cox proportional hazards modeling showed that the expression levels of many PBMC transcripts were predictors for the patient outcomes of time to progression and overall survival (time to death). Supervised class prediction approaches identified multivariate expression patterns in PBMCs capable of assigning favorable outcomes of time to death and time to progression in a test set of renal cancer patients, with overall performance accuracies of 72% and 85%, respectively. CONCLUSIONS: The present study provides the first example of gene expression profiling in peripheral blood, a clinically accessible surrogate tissue, for identifying patterns of gene expression associated with higher likelihoods of positive outcome in patients with a solid tumor.

Adolescent↗
PubMed2005

Transcriptional profiling identifies gene expression changes associated with IFN-alpha tolerance in hepatitis C-related hepatocellular carcinoma cells.

PURPOSE: Treatment with IFN-alpha therapy has been shown to exhibit antitumor effects on patients with hepatocellular carcinoma (HCC). However, individual responses remained unpredictable because of the frequent presence of intrinsic or acquired IFN-alpha resistance. Hence, delineation of molecular targets implicated in the resistant pathway holds value in refining the therapeutic benefits of IFN-alpha. EXPERIMENTAL DESIGN: The current study analyzed the effect of IFN-alpha in human HCC cells. Three hepatitis C virus (HCV)-related, five hepatitis B virus (HBV)-related and two non-B non-C-related cell lines were subjected to IFN-alpha treatment and the cytotoxic effect on cell viability was measured. Further analysis by cDNA microarray and quantitative reverse transcription-PCR were conducted to examine the gene expression changes that mediated the IFN-alpha resistance observed. RESULTS: According to the IC(50) values determined, HCV-related cell lines indicated distinct resistance (IC(50), 389-1468 units/mL) compared with the HBV-related (IC(50), 11-77 units/mL) and non-B non-C-related cell lines (IC(50), 24-108 units/mL). Unsupervised hierarchical clustering on array data indicated three HCV-related cell lines to cluster independently from the sensitive cell lines, suggesting discrete features in association with IFN-alpha tolerance. Moreover, Significance Analysis of Microarrays analysis indicated the differential expression of 149 expressed sequence tags that represented 51 up-regulated and 98 down-regulated genes in the resistant cell lines. Comparing the temporal pattern of gene expression between 6- and 24-hour treatments, candidate genes that were considerably induced with time were further highlighted in the tolerant HCV-related cell lines. These candidates were verified by quantitative reverse transcription-PCR, which confirmed the down-regulation of UBA2, ZNF185, and FOXF1 and up-regulation of UBE4B in the drug-tolerant cells. CONCLUSIONS: Our present study showed that the insensitivity to IFN-alpha therapy in HCC cells is associated with drug-inducible transcriptional alterations. Furthermore, our investigation highlighted potential candidate genes in conferring an anti-apoptotic effect toward IFN-alpha treatment.

Carcinoma, Hepatocellular↗
PubMed2006

Cancer patient flows discovery in DRG databases.

In France, cancer care is evolving to the design of regional networks, so as to coordinate expertise, services and resources allocation. Existing information systems along with data-mining tools can provide better knowledge on the distribution of patient flows. We used one year data of the French Diagnosis Related Groups (DRGs) based system to perform our analysis. Formal Concept Analysis has been used to build Iceberg Lattices of cancer patient flows in the French region of Lorraine. This unsupervised conceptual clustering method allowed us to describe patients flows with an easily understandable visual representation.

Continuity of Patient Care↗
PubMed2007

Gene expression studies of hepatitis virus-induced woodchuck hepatocellular carcinoma in correlation with human results.

The lack of good molecular markers for diagnosis as well as treatment assessment has rendered the hepatocellular carcinoma (HCC) a major challenge in health care. In this study, woodchucks were used as an animal model for hepatitis virus-induced HCC, and gene expression studies were performed using a human oligonucleotide microarray. An analysis approach combing supervised significant analysis of microarray (SAM), prediction analysis of microarray (PAM), and unsupervised hierarchical cluster methodologies statistically determined 211 upregulated and 78 downregulated genes between liver cancer and non-cancer liver tissues, and demonstrated > or = 93% accuracy in classifying the tissue samples. RT-PCR results confirmed the differential expression of selected sequenced woodchuck genes (SAT, IDH3B, SCD) in the microarray. Our study showed that differentially expressed genes were involved in transcription, RNA splicing, translation, cell cycle, metabolism, protein folding and degradation, apoptosis, immune response, metal binding, etc. Interestingly, some genes were involved with signaling pathways such as Ras/MAPK (MAPKAP1), Src-dependent pathways (CSK), hedgehog signaling pathway (HHIP), while Wnt signaling pathway may not be dominant in woodchuck HCC as shown by the downregulation of beta-catenin (TNNB1) and the upregulation of CXXC4 and CSNK2B. Numerous genes found in this study were also differentially expressed in human HCC and many other human cancers including breast, prostate and lung cancers, etc., serving as tumor suppressors, promoters, prognostic markers or chemotherapy targets. In conclusion, this study has demonstrated the robustness of the data analysis and the potential of using human microarrays on woodchuck samples. In particular, some of the differentially expressed genes in the woodchuck HCC can be further explored for possible molecular imaging targets or biological markers in human HCC.

Animals↗
PubMed2006

Unsupervised classification of single particles by cluster tracking in multi-dimensional space.

In cryo-electron microscopy (cryo-EM) single-particle reconstruction, the heterogeneity of two-dimensional projection image data resulting from the co-existence of different conformational or ligand binding states of a macromolecular complex remains a major obstacle as it impairs the validity of reconstructed density maps and limits the progress toward higher resolution. Classification of cryo-EM data according to the different conformations is difficult because of the coexistence of multiple orientations in a single dataset. Here, we present an unsupervised classification method, termed cluster tracking, which utilizes the continuity in multi-dimensional space induced by angular adjacency of projections in large datasets. In a proof of concept, the testing of cluster tracking on simulated projection data, which were generated from multiple conformations and orientations of an existing volume, produced clusters that are consistent with the conformational identity of the data. The application of the method to experimental cryo-EM projection data is found to result in a partition similar to the one generated by supervised classification.

Algorithms↗
PubMed2005

Unsupervised pattern recognition: an introduction to the whys and wherefores of clustering microarray data.

Clustering has become an integral part of microarray data analysis and interpretation. The algorithmic basis of clustering -- the application of unsupervised machine-learning techniques to identify the patterns inherent in a data set -- is well established. This review discusses the biological motivations for and applications of these techniques to integrating gene expression data with other biological information, such as functional annotation, promoter data and proteomic data.

Algorithms↗
PubMed1996

A hybrid approach to clustering biomedical data.

Cluster analysis is often used in an attempt to determine the number and characteristics of patterns present in vectors of biomedical response parameters. An iterative clustering procedure based upon the combined supervised and unsupervised learning algorithm of Pao is described. The present paper illustrates this procedure using growth curves of indices of family functioning in adaptation to pediatric chronic illness. This clustering procedure is based upon neural network approaches to supervised (discriminant analysis) and unsupervised (cluster analysis) learning and is similar to fuzzy set algorithms developed to assess the degree of relatedness among a number of discrete units.

Cluster Analysis↗
PubMed2026

Integrative multi-omics profiling of insomnia-related molecular features reveals microbiome, immune, and therapy-relevant heterogeneity in colorectal cancer.

Emerging evidence implicates insomnia as a potential risk factor in carcinogenesis, potentially involving systemic inflammation, circadian disruption, and microbiome alterations. However, the molecular associations linking insomnia-related features to colorectal cancer (CRC), particularly with respect to tumor biology, immune microenvironmental states, and therapy-relevant phenotypes, remain largely unexplored. Multi-omics integration of genomic, transcriptomic, and microbiome data from 3,026 CRC patients across seven independent cohorts, including a large, well-annotated Clinical Omics study of Colorectal Cancer in China (COCC) cohort, enabled insomnia-based molecular classification through unsupervised non-negative matrix factorization (NMF) clustering. The insomnia subtype (IS) was biologically characterized via pathway enrichment, immune deconvolution, microbial profiling, and single-cell transcriptomics. Furthermore, an insomnia score (ISscore) was developed and validated in multiple cohorts for risk stratification and assessment of treatment-response-related indicators in CRC. Unsupervised clustering revealed two distinct molecular subtypes (IS1/IS2), with IS2 demonstrating significantly poorer survival. IS2 exhibited marked activation of EMT/angiogenesis pathways versus cell cycle activation in IS1. The IS2 microenvironment showed increased immunosuppression-related infiltration and exhausted T cell signatures, together with intratumoral microbiome variation characterized by depletion of Ruminococcaceae UCG-002 and enrichment of Hungatella/Selenomonas. The ISscore system stratified survival risk and was associated with computational indicators of immunotherapy response. Single-cell analysis nominated PPIA-BSG as a potential cell-cell communication signal involving high-ISscore tumor cells, CXCL12+ endothelial cells, and CLEC9A+ dendritic cell subsets. This multi-omics characterization of insomnia-CRC interplay suggests that insomnia-related molecular features are associated with an immunologically distinct and microbiome-altered tumor ecosystem. The ISscore provides a reproducible framework for capturing insomnia-related molecular heterogeneity, supporting risk stratification and future evaluation of therapy-relevant phenotypes.IMPORTANCEChronic insomnia affects millions, but it is not typically considered a cancer risk factor. Our study, analyzing vast biological data from over 3,000 colorectal cancer patients, uncovers a potential link between a person's predisposition to insomnia and their risk of developing this disease. This suggests that the biological pathways related to sleep may play a role in cancer development. Understanding this connection opens up new avenues for identifying individuals at higher risk and developing novel prevention strategies for colorectal cancer.

colorectal cancer↗
PubMed2004

Using unsupervised learning with variational bayesian mixture of factor analysis to identify patterns of glaucomatous visual field defects.

PURPOSE: To determine whether an unsupervised machine learning classifier can identify patterns of visual field loss in standard visual fields consistent with typical patterns learned by decades of human experience. METHODS: Standard perimetry thresholds for 52 locations plus age from one eye of each of 156 patients with glaucomatous optic neuropathy (GON) and 189 eyes of healthy subjects were clustered with an unsupervised machine classifier, variational Bayesian mixture of factor analysis (vbMFA). RESULTS: The vbMFA formed five distinct clusters. Cluster 5 held 186 of 189 fields from normal eyes plus 46 from eyes with GON. These fields were then judged within normal limits by several traditional methods. Each of the other four clusters could be described by the pattern of loss found within it. Cluster 1 (71 GON + 3 normal optic discs) included early, localized defects. A purely diffuse component was rare. Cluster 2 (26 GON) exhibited primarily deep superior hemifield defects, and cluster 3 (10 GON) held deep inferior hemifield defects only or in combination with lesser superior field defects. Cluster 4 (6 GON) showed deep defects in both hemifields. In other words, visual fields within a given cluster had similar patterns of loss that differed from the predominant pattern found in other clusters. The classifier separated the data based solely on the patterns of loss within the fields, without being guided by the diagnosis, placing 98.4% of the healthy eyes within the same cluster and spreading 70.5% of the eyes with GON across the other four clusters, in good agreement with a glaucoma expert and pattern standard deviation. CONCLUSIONS: Without training-based diagnosis (unsupervised learning), the vbMFA identified four important patterns of field loss in eyes with GON in a manner consistent with years of clinical experience.

Algorithms↗
PubMed1998

Dynamic On-line Clustering and State Extraction: An Approach to Symbolic Learning.

Although recurrent neural nets have been moderately successful in learning to emulate finite-state machines (FSMs), the continuous internal state dynamics of a neural net are not well matched to the discrete behavior of an FSM. We describe an architecture, called DOLCE, that allows discrete states to evolve in a net as learning progresses. DOLCE consists of a standard recurrent neural net trained by gradient descent and an adaptive clustering technique that quantizes the state space. We describe two implementations of DOLCE. The first implementation, called DOLCE(u), uses an adaptive clustering scheme in an unsupervised mode to determine both the number of clusters and the partitioning of the state space as learning progresses. The second model, DOLCE(s), uses a Gaussian Mixture Model in a supervised learning framework to infer the states of an FSM. DOLCE(s) is based on the assumption that a finite set of discrete internal states is required for the task, and that the actual network state belongs to this set but has been corrupted by noise due to inaccuracy in the weights. DOLCE(s) learns to recover the discrete state with maximum a posteriori probability from the noisy state. Simulations show that both implementations of DOLCE lead to a significant improvement in generalization performance over earlier neural net approaches to FSM induction. The idea of adaptive quantization is not just applicable to DOLCE but can be applied to other domains as well.

Journal Article↗
PubMed2006

Unsupervised classification of ventricular extrasystoles using bounded clustering algorithms and morphology matching.

Ventricular extrasystoles (VE) are ectopic heartbeats involving irregularities in the heart rhythm. VEs arise in response to impulses generated in some part of the heart different from the sinoatrial node. These are caused by the premature discharge of a ventricular ectopic focus. VEs after myocardial infarction are associated with increased mortality. Screening of VEs is typically a manual and time consuming task that involves analysis of the heartbeat morphology, QRS duration, and variations of the RR intervals using long-term electrocardiograms. We describe a novel algorithm to perform automatic classification of VEs and report the results of our validation study. The proposed algorithm makes use of bounded clustering algorithms, morphology matching, and RR interval length to perform automatic VE classification without prior knowledge of the number of classes and heartbeat features. Additionally, the proposed algorithm does not need a training set.

Algorithms↗
PubMed1997

A fuzzy clustering approach to EP estimation.

The problem of extracting a useful signal (a response) buried in relatively high amplitude noise has been investigated, under the conditions of low signal-to-noise ratio. In particular, we present a method for detecting the "true" response of the brain resulting from repeated auditory stimulation, based on selective averaging of single-trial evoked potentials. Selective averaging is accomplished in two steps. First, an unsupervised fuzzy-clustering algorithm is employed to identify groups of trials with similar characteristics, using a performance index as an optimization criterion. Then, typical responses are obtained by ensemble averaging of all trials in the same group. Similarity among the resulting estimates is quantified through a synchronization measure, which accounts for the percentage of time that the estimates are in phase. The performance of the classifier is evaluated with synthetic signals of known characteristics, and its usefulness is demonstrated with real electrophysiological data obtained from normal volunteers.

Algorithms↗