Clustering is the unsupervised classification of patterns (observations, data items, or feature vectors) into groups (clusters). The clustering problem has been addressed in many contexts and by researchers in many disciplines; this reflects its broad appeal and usefulness as one of the steps in exploratory data analysis. However, clustering is a difficult problem combinatorially, and differences in assumptions and contexts in different communities has made the transfer of useful generic concepts and methodologies slow to occur. This paper presents an overview of pattern clustering methods from a statistical pattern recognition perspective, with a goal of providing useful advice and references to fundamental concepts accessible to the broad community of clustering practitioners. We present a taxonomy of clustering techniques, and identify cross-cutting themes and recent advances. We also describe some important applications of clustering algorithms such as image segmentation, object recognition, and information retrieval.

/pdf/data-clustering-a-review-4c452f5cab.pdf

Data clustering: a review

2016 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure: The Task Force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC). Developed with the special contribution of the Heart Failure Association (HFA) of the ESC.

https://caltcm.memberclicks.net/assets/documents/acc_aha_chf_2013_guidelines.pdf

2013 ACCF/AHA Guideline for the Management of Heart Failure: A Report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines

Atherosclerosis, the principal cause of heart attack, stroke and gangrene of the extremities, is responsible for 50% of all mortality in the USA, Europe and Japan. The lesions result from an excessive, inflammatory-fibroproliferative response to various forms of insult to the endothelium and smooth muscle of the artery wall. A large number of growth factors, cytokines and vasoregulatory molecules participate in this process. Our ability to control the expression of genes encoding these molecules and to target specific cell types provides opportunities to develop new diagnostic and therapeutic agents to induce the regression of the lesions and, possibly, to prevent their formation.

The pathogenesis of atherosclerosis: a perspective for the 1990s

The Adult Treatment Panel III (ATP III) of the National Cholesterol Education Program issued an evidence-based set of guidelines on cholesterol management in 2001. Since the publication of ATP III, 5 major clinical trials of statin therapy with clinical end points have been published. These trials addressed issues that were not examined in previous clinical trials of cholesterol-lowering therapy. The present document reviews the results of these recent trials and assesses their implications for cholesterol management. Therapeutic lifestyle changes (TLC) remain an essential modality in clinical management. The trials confirm the benefit of cholesterol-lowering therapy in high-risk patients and support the ATP III treatment goal of low-density lipoprotein cholesterol (LDL-C) <100 mg/dL. They support the inclusion of patients with diabetes in the high-risk category and confirm the benefits of LDL-lowering therapy in these patients. They further confirm that older persons benefit from therapeutic lowering of LDL-C. The major recommendations for modifications to footnote the ATP III treatment algorithm are the following. In high-risk persons, the recommended LDL-C goal is <100 mg/dL, but when risk is very high, an LDL-C goal of <70 mg/dL is a therapeutic option, ie, a reasonable clinical strategy, on the basis of available clinical trial evidence. This therapeutic option extends also to patients at very high risk who have a baseline LDL-C <100 mg/dL. Moreover, when a high-risk patient has high triglycerides or low high-density lipoprotein cholesterol (HDL-C), consideration can be given to combining a fibrate or nicotinic acid with an LDL-lowering drug. For moderately high-risk persons (2+ risk factors and 10-year risk 10% to 20%), the recommended LDL-C goal is <130 mg/dL, but an LDL-C goal <100 mg/dL is a therapeutic option on the basis of recent trial evidence. The latter option extends also to moderately high-risk persons with a baseline LDL-C of 100 to 129 mg/dL. When LDL-lowering drug therapy is employed in high-risk or moderately high-risk persons, it is advised that intensity of therapy be sufficient to achieve at least a 30% to 40% reduction in LDL-C levels. Moreover, any person at high risk or moderately high risk who has lifestyle-related risk factors (eg, obesity, physical inactivity, elevated triglycerides, low HDL-C, or metabolic syndrome) is a candidate for TLC to modify these risk factors regardless of LDL-C level. Finally, for people in lower-risk categories, recent clinical trials do not modify the goals and cutpoints of therapy.

/pdf/implications-of-recent-clinical-trials-for-the-national-va66u0my8e.pdf

Implications of Recent Clinical Trials for the National Cholesterol Education Program Adult Treatment Panel III Guidelines

Multichannel integrated action-potential recordings (called electromyometrograms or EMMGs) from 10 individuals with various functional symptoms and disorders were selected for statistical examination to obtain evidence bearing on the topic in question Each EMMG recorded the action-potential levels from four regions simultaneously via four channels of integrating differential amplifiers The subjects sat quietly, awake and alert, with eyes open and head erect Electrodes were placed over antagonistic muscle groups in the following regions: (1) forehead, (2) jaw-throat, (3) right forearm, (4) left leg Successive 1-minute action-potential levels from each of these regions were used to calculate within-subject correlations (r) and predictabilities (r
2 andR
2) among the different regions Mean and median correlations (r) and predictabilities (r
2 andR
2) for each of the following pairs of regions were found to be weak: (1) forehead and jaw-throat, (2) forehead and forearm, (3) forehead and leg, (4) jaw-throat and forearm, (5) jaw-throat and leg, (6) forearm and leg The findings support the statement that the frontalis or any of the other regions measured are poor indicators of the amount of activity elsewhere in the skeletal musculature of awake and alert individuals When combined with basic principles of neuroanatomy and neurophysiology, the findings also support the statement that there is probably no portion of the skeletal musculature that can be used to quantify the activity in other portions

Is frontalis activity a reliable indicator of the activity in other skeletal muscles

B-scan ultrasonic measurements of lumbar spinal canal diameter were examined as predictors of industrial back pain complaints and extended work loss. Baseline data were collected on 3,020 Washington State aircraft manufacturing workers, and over a mean 3.7-year follow-up period 352 subjects reported industrial back pain complaints. Mean canal measurements of subjects with industrial back pain complaints were smaller at all spinal levels than in subjects without complaints. The mean differences between the groups, however, were extremely small (0.07 mm to 0.51 mm), and not all levels were statistically significant. The relative risk for an L5-S1 measurement 2 standard deviations below the mean was 1.4, yet the measurement explained less than 1% of the uncertainty in predicting complaints. No association was found between canal measurements and claims with extended work loss of greater than one month. The imprecision of the measurements and poor predictive ability indicate that B-scan ultrasonography, as used in this study, is of dubious screening value.

B-scan ultrasonic measurement of the lumbar spinal canal as a predictor of industrial back pain complaints and extended work loss.

Traditionally, intravenous (IV) drugs in decompensated heart failure (HF) have been evaluated based on their hemodynamic effects. Thus, no endpoint to measure symptomatic improvement in decompensated HF has yet been validated. To do so, we adapted the clinical composite (which has been validated in trials of chronic HF) to the evaluation of IV levosimendan in a pilot trial of acute decompensated HF (REVIVE-1).

A total of 100 patients who were hospitalized for worsening HF and had dyspnea at rest despite IV diuretics were randomized (double-blind) to receive placebo (PBO) (n = 49) or IV levosimendan (LS) (n = 51), given as a loading dose of 12 μg/kg over 10 min and followed by a continuous infusion (0.1 μg/kg/hour for 50 min and 0.2 μg/kg/hour for 23 hours); patients were then followed closely for an additional 4 days. Patients were classified as improved if they reported their HF to be moderately or markedly improved at specific time points. Patients were classified as worse if they died or received an IV medication for worsening HF during the 5-day study period. Patients were considered unchanged if they were neither improved nor worse. Different models using different definitions for improvement and worsening were prospectively and retrospectively examined.

The initial model defined improvement based on the responses at 24 hours and at 5 days and restricted the definition of worsening to the use of IV vasodilators or inotropic agents. Using this model, improvement was observed more frequently with LS than with PBO (49% vs 33%), with no between-group difference in the proportion of patients who were worse (overall P = 0.229). When the definition of worsening was expanded to include the use of IV diuretics for worsening HF and confined to the occurrence of clinical events, LS-treated patients not only were more likely to show improvement (51% vs 33%) but were less likely to exhibit worsening (20% vs 35%) (overall P = 0.043). When the definition of improvement was expanded to include the responses at 6 hours (in addition to 24 hours and 5 days), the separation between the treatment groups increased even further (improvement in 33% vs 14% and worsening in 24% vs 37%) (overall P = 0.029) (LS vs PBO). The greater level of improvement in the clinical composite on levosimendan was supported by significant reductions in median plasma B-type natriuretic peptide concentrations compared with placebo at both 24 hours (-303 pg/ml vs -77.5 pg/ml, P = 0.001) and 5 days (-376 pg/ml vs -99.5 pg/ml, P = 0.027).

These findings indicate that a clinical composite approach can be used to develop an endpoint that distinguishes the effects of IV LS in the setting of acute decompensated HF. The new endpoint is now being used prospectively to evaluate the effects of LS in a definitive second study (REVIVE-2).

/pdf/development-of-a-comprehensive-new-endpoint-for-the-2xjt8iemrl.pdf

Development of a comprehensive new endpoint for the evaluation of new treatments for acute decompensated heart failure: results with levosimendan in the REVIVE 1 study

Nitroglycerin improves the exercise performance of patients with coronary heart disease. A retrospective study of data collected while monitoring 103 such patients was focused on the various types of arrhythmia and the effects of nitroglycerin on the frequency of occurrence. After nitroglycerin administration, occurrences were fewer. This applied more to angina patients than to those with healed myocardial infarction.

Exertional arrhythmia and nitgroglycerin.

The identification problem of concern here is the estimation of a real function $f(x)$ by means of noisy observations $\{ (X_i ,f(X_i ) + \eta _i (X_i ))\} $ of its pairs, the $X_i $’s being chosen independently according to some fixed law P. The approach taken for estimation is the “potential function” method (its sources are referenced herein), to wit: Choose $f_0 $ arbitrarily and define the sequence $\{ f_n \} $ by the recursive relation $f_{n + 1} (x) = f_n (x) + \gamma _n (f(X_{n + 1} ) + \eta (X_{n + 1} ) - f_n (X_{n + 1} ))K(X_{n + 1} ,x)$, K being a positive symmetric kernel. From earlier publications it is known that under certain mild restrictions $E[\| {f_n - f} \|^2 ] \to 0$ in the $L_2 (p)$-norm. Rates of convergence have been obtained in the restrictive case that $K(x,y) = \sum_{i - 1}^N {\lambda _i^2 } \phi _i (x)\phi _i (y)$ and $f(x) \in {\operatorname{span}}\{ \phi _i ,1 \leqq i \leqq N\} $. The contribution of this paper is to prove that while no uniform bounds exist in the $L_2 (p)$-n...

Lloyd D. Fisher

Papers

Is frontalis activity a reliable indicator of the activity in other skeletal muscles

B-scan ultrasonic measurement of the lumbar spinal canal as a predictor of industrial back pain complaints and extended work loss.

Development of a comprehensive new endpoint for the evaluation of new treatments for acute decompensated heart failure: results with levosimendan in the REVIVE 1 study

Exertional arrhythmia and nitgroglycerin.

Uniform Convergence of the Potential Function Algorithm