SEARCH · PubMed Health

Results for “Classification Algorithms”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

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

At least 487 records · Page 27Linked to original sources

PubMed1993

A neural network classifier in experimental particle physics.

A classification problem in high energy physics has been solved on simulated data using a simple multilayer perceptron comprising binary units which was trained with the CHIR algorithm. The unstable training of such a network on a nonseparable set has been overcome by selecting those weight vectors with good performance while providing a flexible choice of the two types of classification errors. Specific features of the problem have been exploited in order to simplify and optimize the solution which has been compared to the popular backpropagation algorithm and found to perform on a similar level. Additional aspects of this work are the use of the CHIR algorithm on continuous input and incorporating the classic idea of a phi-machine in a multilayer perceptron.

Algorithms↗

PubMed1994

Hidden Markov models of biological primary sequence information.

Hidden Markov model (HMM) techniques are used to model families of biological sequences. A smooth and convergent algorithm is introduced to iteratively adapt the transition and emission parameters of the models from the examples in a given family. The HMM approach is applied to three protein families: globins, immunoglobulins, and kinases. In all cases, the models derived capture the important statistical characteristics of the family and can be used for a number of tasks, including multiple alignments, motif detection, and classification. For K sequences of average length N, this approach yields an effective multiple-alignment algorithm which requires O(KN2) operations, linear in the number of sequences.

Algorithms↗

PubMed1994

Classification of impulse radar waveforms using neural networks.

In this paper, it is demonstrated that multilayer neural networks, trained with the backpropagation algorithm and radial basis functions, can classify impulse radar waveforms from three different asphalt-covered bridge decks, each with its own structure. It might be thought that the thickness of asphalt and the depth of concrete over the reinforcing bars would be nearly constant for any one bridge deck; however in practice this is not the case. There are often significant changes in the thickness of the asphalt and the cover over reinforcement. Furthermore, a certain amount of damage to the concrete caused by severe winter climate often produces a random variation in the reflected waveforms obtained from different locations. These factors lead to a significant number of combinations of waveforms that can be obtained from any given structural type of deck. The classification accuracies achieved ranged between 89.9% and 100%. The accuracies achieved after using principal components analysis to reduce the dimensionality of the input data ranged between 95.6% and 100%.

Algorithms↗

PubMed1998

Algorithms for radiological image registration and their clinical application.

This paper reviews recent work in radiological image registration and provides a classification of image registration by type of transformation and by methods employed to compute the transformation. The former includes transformation of 2D images to 2D images of the same individual, transformation of 3D images to 3D images of the same individual, transformation of images to an atlas or model, transformation of images acquired from a number of individuals, transformations for image guided interventions including 2D to 3D registration and finally tissue deformation in image guided interventions. Recent work on computing transformations for registration using corresponding landmark based registration, surface based registration and voxel similarity measures, including entropy based measures, are reviewed and compared. Recently fully automated algorithms based on voxel similarity measures and, in particular, mutual information have been shown to be accurate and robust at registering images of the head when the rigid body assumption is valid. Two approaches to modelling soft tissue deformation for applications in image guided interventions are described. Validation of complex processing tasks such as image registration is vital if these algorithms are to be used in clinical practice. Three alternative validation strategies are presented. These methods are finding application outside the original domain of radiological imaging.

Algorithms↗

PubMed1991

Computer-assisted statistical analyses of enzyme and ribosomal DNA electrophoretic polymorphism in Yersinia.

The intra- and inter-species differentiation of 90 strains of Yersinia belonging to six species were studied independently by computer-assisted statistical analysis of data from enzyme electrophoretic polymorphism and ribosomal DNA (rDNA) restriction fragment length polymorphism. Two correspondence analyses (CA) demonstrated the concordance between the bacterial classification obtained from enzymatic and genomic data. This concordance was reinforced by an algorithm of the correspondence established between the two dendrograms drawn from previous computations. Comparison of CA with similarity analysis (also called "numerical taxonomy") indicated that the intra- and inter-species differentiation obtained by the two methods are similar. The advantage of CA is that it gives a synthetic geometrical representation of the results (factorial planes), displaying both the main features of the clusters of strains (location and dispersion) and their essential character (i.e.: enzyme electrophoretic variant, rDNA fragment size).

Algorithms↗

PubMed1985

Computerized EEG pattern classification by adaptive segmentation and probability density function classification. Clinical evaluation.

A series of 63 clinical EEGs showing a variety of normal and abnormal patterns was analysed by computer with particular reference to the different types of pattern within the same EEG. Boundaries between different patterns were established by means of adaptive segmentation, so that the duration of the resulting segments was determined by the particular EEG itself (thus the term 'adaptive'). Four channels from each EEG were analysed, paired (left and right) channels were simultaneously segmented and analysed interactively. Similar segments were then clustered without supervision by estimating a probability density function in a 2-dimensional 'feature space' having dimensions of mean frequency and mean power. Individual clusters emerged as well-defined peaks of the surface, individual segments or small groups of duration insufficient to constitute a separate cluster, being identified as 'singular events' (e.g., rare sharp waves, artifacts). The autocorrelation function was used to characterize the EEG both for the segmentation and for the subsequent clustering of the resulting segments. In confirmation of our previous work, adaptive segmentation based on the autocorrelation function of the EEG was found to be quite satisfactory. Unsupervised clustering by estimation of the probability density function in feature space was found to give the correct number of clusters (usually less than 5) in a majority of the records (65%), but in the remaining minority of cases (35%), either overclustering or underclustering occurred. Further, the 'singular events' were occasionally partly included in a formal cluster. Comparison of these results of EEG clustering by unsupervised probability density function estimation with earlier results obtained by supervised hierarchical clustering suggests that there may be subtle cues used by the electroencephalographer in the classification of EEG patterns which have not been adequately approximated by the computer algorithms thus far used in this work. Hence at least some minimal degree of supervision in the clustering process may be necessary, at least for the present. On the other hand, the method recommends itself for the representation of illustrative EEG summaries which, in conjunction with a short written report, would provide the clinical neurologist with a sufficient picture of the real EEG without, in most cases, the need to inspect the original record.

Action Potentials↗

PubMedDate not supplied

Automatic identification of significant graphoelements in multichannel EEG recordings by adaptive segmentation and fuzzy clustering.

A new approach to visual evaluation of long-term EEG recordings is proposed. The method is based on multichannel adaptive segmentation, subsequent feature extraction, automatic classification of the acquired segments by fuzzy cluster analysis (fuzzy c-means algorithm), and on the distinguishing of thus identified EEG segments by colour directly in the EEG record. The black and white variant of the described automatic system is presented. The method was evaluated by applying it to simulated artificial data and to real EEG recordings; some of the illustrative results are shown. In addition, the performance of this system is evaluated and the first experience with its application to routine EEG recordings is discussed.

Algorithms↗

PubMed1989

Heuristic potency of the minimum spanning tree (MST) method in toxicology.

A rapid and manual mathematical method for comparing ecotoxicologic data has been investigated. It uses a classification procedure based on the calculation of chi 2 distances between the different elements of a data matrix. The classification is carried out using a simple graph-theoretical procedure (Kruskal's algorithm), allowing the construction of minimum spanning tree (MST). From an ecotoxicologic data base of 18 bacterial tests carried out on eight heavy metals the construction of the MST is explained in detail. Even if the results obtained are directly dependent on the data set chosen, they illustrate the heuristic potency of the minimum spanning tree method in comparing and estimating the dependence between environmental data.

Bacteria↗

PubMed1982

Computer based pattern recognition of carotid arterial disease using pulsed Doppler ultrasound.

A minicomputer based system has been developed for studying carotid artery blood flow data obtained for a combined B-mode, pulsed Doppler ultrasound scanner. The goals of this work are to devise and improve techniques for estimating the extent of atherosclerosis at the carotid artery bifurcation. Features are automatically extracted from spectrum analyzed Doppler blood flow data. Five statistical pattern recognition algorithms are compared, with cross validation being used to improve the estimate of classification accuracy. A data collection protocol has been devised in which four sites are studied along each carotid arterial system. Classification of unknowns is done using a hierarchy of three decisions.

Adult↗

PubMed1996

The need for pediatric-specific triage criteria: results from the Florida Trauma Triage Study.

OBJECTIVE: The objective of the Florida Trauma Triage Study was to assess the performance of state-adopted field triage criteria. The study addressed three specific age groups: pediatric (age < 15 years), adult (age 15-54 years), and geriatric (age 55+ years). Since 1990, Florida has used a uniform set of eight triage criteria, known as the trauma scorecard, for triaging adult trauma patients to state-approved trauma centers. However, only five of the criteria are recommended for use with pediatric patients. This article presents the findings regarding the performance of the scorecard when applied to a pediatric population. DESIGN: We used state trauma registry data linked to state hospital discharge data in a retrospective analysis of trauma patients transported by prehospital providers to any acute care hospital within nine selected Florida counties between July 1, 1991, and December 31, 1991. We used cross-table and logistic regression analysis to determine the ability of triage criteria to correctly identify patients who were retrospectively defined as major trauma. We applied the field criteria to physiologic and anatomy/mechanism of injury data contained in the trauma registry to "score" the patient as major or minor trauma. To make our retrospective determination of major or minor trauma we used the protocols developed by an expert medical panel as described by E. J. MacKenzie et al. (1990). MAIN OUTCOME MEASURES: We calculated sensitivity, specificity, and the corresponding over- and undertriage rates by comparing patient classifications (major or minor trauma) produced by the triage criteria and the retrospective algorithm. We used logistic regression to identify which triage criteria were statistically significant in predicting major trauma. RESULTS: Pediatric cases accounted for 9.2% of the total study population, 6.0% of all hospitalized cases, and 6.8% of all trauma deaths. Of the 1505 pediatric cases available for analysis, the triage criteria classified 269 cases as expected major trauma and 1236 cases as expected minor trauma. The retrospective algorithm classified 78 cases as expected major trauma and 1427 cases as expected minor trauma. The resulting specificity is 84.8% (15.2% overtriage), and the sensitivity is 66.7% (33.3% undertriage). Logistic regression indicated that, of the eight state-adopted field triage criteria, only the Glasgow coma score, ejection from vehicle, and penetrating injuries have a statistically significant impact on predicting major trauma in pediatric patients. CONCLUSIONS: Although the state-adopted trauma scorecard, applied to a pediatric population, produced acceptable overtriage, it did not produce acceptable undertriage. However, our undertriage rate is comparable to the results of other published studies on pediatric trauma. As a result of the Florida Trauma Triage Study, a new pediatric triage instrument was developed. It is currently being field-tested.

Adolescent↗

PubMed1996

Infrared spectroscopy of dystrophic mdx mouse muscle tissue distinguishes among treatment groups.

Four groups of mdx mice (deflazacort, high dose of 1.5 mg/kg and low dose of 0.75 mg/kg; prednisone, 1.0 mg/kg; and a placebo) were examined in a double-blind protocol. The experiments tested the hypothesis that infrared spectroscopy can distinguish among gastrocnemius muscle tissues derived from dystrophic animals (n = 22) from different treatment groups and from control muscle tissue (n = 23). Results showed that muscle, inflamed muscle, and tendon can be distinguished on the basis of their infrared absorption patterns. Distinctions among the spectra of the four treatment groups were sought with automated pattern-recognition methods. These classification methods, based either on spectral regions (900-1,500 cm-1) or on principal-component analysis, were in close agreement, assigning 15 or 16, respectively, of 22 mdx spectra to the correct treatment group. Both trials cleanly separated the high-dose deflazacort from the placebo group of muscles, whereas the prednisone and low-dose deflazacort groups were persistently confused in these classifications. Changes in the histology of muscle inflammation paralleled the spectral-classification results. Thus the proposed method, combining infrared spectroscopy with pattern-recognition algorithms, can distinguish treatment effects on muscle tissue. Specific spectral features characteristic of tissue type, disease progression, and treatment effects are not yet elucidated.

Algorithms↗

PubMed1985

Imaging approach to the suspected renal mass.

The authors present their algorithmic approach to the detection, characterization, and staging of renal masses. Based on classification of urographic findings, the patient may be triaged to the appropriate cross-sectional or invasive imaging modality that will result in the most cost-effective management.

Abscess↗

PubMed1987

[Staging of pulmonary cancer, establishment of M1].

The rapid and widespread development of imaging techniques during the last decade has markedly modified the previous algorithms used in the staging of pulmonary carcinoma, particularly M0/M1 in the TNM classification and the directives of the American Thoracic Society. Sensitivity and specificity of each method are reviewed according to the most frequent metastatic sites of bronchopulmonary carcinoma. Presently, CT is the most efficient technique for detection and display of metastases of the contralateral lung, brain, adrenal glands and retroperitoneal lymph nodes. Ultrasound is equal or even slightly superior to CT for the detection of liver metastases. The superiority of magnetic resonance imaging (MRI) over CT in the detection of brain metastases has already been demonstrated. The results of MRI using fast sequences have recently been demonstrated for imaging of thoracic, abdominal and bone metastases, but confirmation of these first results by prospective studies is needed. Skeletal survey is still obtained by radioisotope scanning.

Bone Neoplasms↗

PubMed1981

Metabolic abnormalities associated with diabetes mellitus, as investigated by gas chromatography and pattern-recognition analysis of profiles of volatile metabolites.

Patterns of volatile metabolites in urine, as obtained by glass-capillary gas chromatography, were investigated by use of a nonparametric pattern-recognition method, in an effort to detect abnormalities associated with diabetes. We used threshold logic unit analysis on a data set consisting of normal subjects and those with diabetes mellitus, and could predict patterns for volatile metabolites as belonging to the proper class in 94.83% of the cases examined. In addition, a feature-extraction algorithm isolated those volatile constituents that are most useful in making the normal/diabetic classification. We used gas chromatography/mass spectrometry to identify important profile constituents. Finally, these same pattern-recognition methods indicated strong sex-related patterns in these volatiles.

Chromatography, Gas↗

PubMed1999

Quantitative follow-up of patients with multiple sclerosis using MRI: technical aspects.

A highly reproducible automated procedure for quantitative analysis of serial brain magnetic resonance (MR) images was developed for use in patients with multiple sclerosis (MS). The intracranial cavity (ICC) was identified on standard dual-echo spin-echo brain MR images using a supervised automated procedure. MR images obtained from one MS patient at 24 time points in the course of a 1-year follow-up were aligned with the images of one of the time points. Next, the contents of the ICC in each MR exam were segmented into four tissues, using a self-adaptive statistical algorithm. Misclassifications due to partial voluming were corrected using a combination of morphologic operators and connectivity criteria. Finally, a connectivity detection algorithm was used to separate the tissue classified as lesions into individual entities. Registration, classification of the contents of the ICC, and identification of individual lesions are fully automatic. Only identification of the ICC requires operator interaction. In each MR exam, the program estimated volumes for the ICC, gray matter (GM), white matter (WM), white matter lesions (WML), and cerebrospinal fluid (CSF). The reproducibility of the system was superior to that of supervised segmentation, as evidenced by the coefficient of variation: CSF supervised 45.9% vs. automated 7.7%, GM 16.0% vs. 1.4%, WM 15.7% vs. 1.3%, and WML 39.5% vs 52.0%. Our results demonstrate that this computerized procedure allows routine reproducible quantitative analysis of large serial MRI data sets.

Adult↗

PubMed1996

[Usefulness of headache classification in the planning of radiological studies].

Neurologists can only classify headaches, not patients. Doing that, they are well aware that this is partially arbitrary. But classification is essential, especially for primary headaches (migraine, tension type headache, cluster headache) to achieve epidemiological and physiopathological studies. The criteria of definition for each type of headache must be established with pertinence. Neurologists, general practitioners as well as radiologists can then use these criteria to imagine an algorithm for the diagnostic procedure of headache patients. The aim of this publication is to describe the IHS classification as well as some common clinical situations (migraine, subarachnoidal haemorrhage, brain tumor, mild head injury.

Brain Neoplasms↗

PubMed1997

Learning systems in biosignal analysis.

In biosignal analysis, the utility of artificial neural networks (ANN) in classifying electromyographic (EMG) data trained with the momentum back propagation algorithm has recently been demonstrated. In the current study, the self-organizing feature map algorithm, the genetics-based machine learning (GBML) paradigm, and the K-means nearest neighbour clustering algorithm are applied on the same set of data. The aim of this exercise is to show how these three paradigms can be used in practice, given that their diagnostic performance is problem- and parameter-dependent. A total of 720 macro EMG recordings were carried out from four groups, from seven normal, nine motor neuron disease, 14 Becker's muscular dystrophy, and six spinal muscular atrophy subjects, respectively. Twenty-three of the subjects were used for training and 13 for evaluating the various models. For each subject, the mean and the standard deviation of the parameters (i) amplitude, (ii) area, (iii) average power and (iv) duration were extracted. The feature vector was structured in two different ways for input to the models: an eight-input feature vector that consisted of both the mean and the standard deviation of the four parameters measured, and a four-input feature vector that included only the mean of the parameters. Also, due to the heterogenous nature of the spinal muscular atrophy group, three class models that excluded this group were investigated. In general, self-organizing feature map and GBML models resulted in comparable diagnostic performance of the order of 80-90% correct classifications (CCs) score for the evaluation set, whereas the K-means nearest neighbour algorithm models gave lower percentage CCs. Furthermore, for all three learning paradigms: better diagnostic performance was obtained for the three class models compared with the four class models; similar diagnostic performance was obtained for both the eight- and four-input feature vectors. Finally, it is claimed that the proposed methodology followed in this work can be applied for the development of diagnostic systems in the analysis of biosignals.

Algorithms↗

PubMed1998

Approaches to the automatic discovery of patterns in biosequences.

This paper surveys approaches to the discovery of patterns in biosequences and places these approaches within a formal framework that systematises the types of patterns and the discovery algorithms. Patterns with expressive power in the class of regular languages are considered, and a classification of pattern languages in this class is developed, covering the patterns that are the most frequently used in molecular bioinformatics. A formulation is given of the problem of the automatic discovery of such patterns from a set of sequences, and an analysis is presented of the ways in which an assessment can be made of the significance of the discovered patterns. It is shown that the problem is related to problems studied in the field of machine learning. The major part of this paper comprises a review of a number of existing methods developed to solve the problem and how these relate to each other, focusing on the algorithms underlying the approaches. A comparison is given of the algorithms, and examples are given of patterns that have been discovered using the different methods.

Algorithms↗