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 811 records · Page 45Linked to original sources

PubMed2004

Classification of anatomical structures in MR brain images using fuzzy parameters.

We present an algorithm that automatically segments and classifies the brain structures in a set of magnetic resonance (MR) brain images using expert information contained in a small subset of the image set. The algorithm is intended to do the segmentation and classification tasks mimicking the way a human expert would reason. The algorithm uses a knowledge base taken from a small subset of semiautomatically classified images that is combined with a set of fuzzy indexes that capture the experience and expectation a human expert uses during recognition tasks. The fuzzy indexes are tissue specific and spatial specific, in order to consider the biological variations in the tissues and the acquisition inhomogeneities through the image set. The brain structures are segmented and classified one at a time. For each brain structure the algorithm needs one semiautomatically classified image and makes one pass through the image set. The algorithm uses low-level image processing techniques on a pixel basis for the segmentations, then validates or corrects the segmentations, and makes the final classification decision using higher level criteria measured by the set of fuzzy indexes. We use single-echo MR images because of their high volumetric resolution; but even though we are working with only one image per brain slice, we have multiple sources of information on each pixel: absolute and relative positions in the image, gray level value, statistics of the pixel and its three-dimensional neighborhood and relation to its counterpart pixels in adjacent images. We have validated our algorithm for ease of use and precision both with clinical experts and with measurable error indexes over a Brainweb simulated MR set.

Algorithms↗

PubMed2003

Identification of bacterial spores using statistical analysis of Fourier transform infrared photoacoustic spectroscopy data.

Fourier transform infrared photoacoustic spectroscopy (FTIR-PAS) has been applied for the first time to the identification and speciation of bacterial spores. A total of forty specimens representing five strains of Bacillus spores (Bacillus subtilis ATCC 49760, Bacillus atrophaeus ATCC 49337, Bacillus subtilis 6051, Bacillus thuringiensis subsp. kurstaki, and Bacillus globigii Dugway) were analyzed. Spores were deposited, with minimal preparation, into the photoacoustic sample cup and their spectra recorded. Principal component analysis (PCA), classification and regression trees (CART), and Mahalanobis distance calculations were used on this spectral library to develop algorithms for step-wise classification at three levels: (1) bacterial/nonbacterial, (2) membership within the spore library, and (3) bacterial strain. Internal cross-validation studies on library spectra yielded classification success rates of 87% or better at each of these three levels. Analysis of fifteen blind samples, which included five samples of spores already in the spectral library, two samples of closely related Bacillus globigii 01 spores not in the library, and eight samples of nonbacterial materials, yielded 100% accuracy in distinguishing among bacterial/nonbacterial samples, membership in the library, and bacterial strains within the library.

Acoustics↗

PubMed2000

Automatic segmentation of lung fields in chest radiographs.

The delineation of important structures in chest radiographs is an essential preprocessing step in order to automatically analyze these images, e.g., for tuberculosis screening support or in computer assisted diagnosis. We present algorithms for the automatic segmentation of lung fields in chest radiographs. We compare several segmentation techniques: a matching approach; pixel classifiers based on several combinations of features; a new rule-based scheme that detects lung contours using a general framework for the detection of oriented edges and ridges in images; and a hybrid scheme. Each approach is discussed and the performance of nine systems is compared with interobserver variability and results available from the literature. The best performance is obtained by the hybrid scheme that combines the rule-based segmentation algorithm with a pixel classification approach. The combinations of two complementary techniques leads to robust performance; the accuracy is above 94% for all 115 images in the test set. The average accuracy of the scheme is 0.969 +/- 0.0080, which is close to the interobserver variability of 0.984 +/- 0.0048. The methods are fast, and implemented on a standard PC platform.

Algorithms↗

PubMed2001

Application of neural networks to the classification of pancreatic intraductal proliferative lesions.

The aim of the study was to test applycability of neural networks to classification of pancreatic intraductal proliferative lesions basing on nuclear features, especially chromatin texture. Material for the study was obtained from patients operated on for pancreatic cancer, chronic pancreatitis and other tumours requiring pancreatic resection. Intraductal lesions were classified as low and high grade as previously described. The image analysis system consisted of a microscope, CCD camera combined with a PC and AnalySIS v. 2.11 software. The following texture characteristics were measured: variance of grey levels, features extracted from the grey levels correlation matrix and mean values, variance and standard deviation of the energy obtained from Laws matrices. Furthermore we used moments derived invariants and basic geometric data such as surface area, the minimum and maximum diameter and shape factor. The sets of data were randomly divided into training and testing groups. The training of the network using the back-propagation algorithm, and the final classification of data was carried out with a neural network simulator SNNS v. 4.1. We studied the efficacy of networks containing from one to three hidden layers. Using the best network, containing three hidden layers, the rate of correct classification of nuclei was 73%, and the rate of misdiagnosis was 3%; in 24% the network response was ambiguous. The present findings may serve as a starting point in search for methods facilitating early diagnosis of ductal pancreatic carcinoma.

Adenocarcinoma↗

PubMed2005

Classification of cancer types by measuring variants of host response proteins using SELDI serum assays.

Protein expression profiling has been increasingly used to discover and characterize biomarkers that can be used for diagnostic, prognostic or therapeutic purposes. Most proteomic studies published to date have identified relatively abundant host response proteins as candidate biomarkers, which are often dismissed because of an apparent lack of specificity. We demonstrate that 2 host response proteins previously identified as candidate markers for early stage ovarian cancer, transthyretin and inter-alpha trypsin inhibitor heavy chain 4 (ITIH4), are posttranslationally modified. These modifications include proteolytic truncation, cysteinylation and glutathionylation. Assays using Surface Enhanced Laser Desorption/Ionization Time of Flight Mass Spectrometry (SELDI-TOF-MS) may provide a means to confer specificity to these proteins because of their ability to detect and quantitate multiple posttranslationally modified forms of these proteins in a single assay. Quantitative measurements of these modifications using chromatographic and antibody-based ProteinChip array assays reveal that these posttranslational modifications occur to different extents in different cancers and that multivariate analysis permits the derivation of algorithms to improve the classification of these cancers. We have termed this process host response protein amplification cascade (HRPAC), since the process of synthesis, posttranslational modification and metabolism of host response proteins amplifies the signal of potentially low-abundant biologically active disease markers such as enzymes.

Algorithms↗

PubMed2006

Classification of pediatric lumbosacral spondylolisthesis.

A surgical classification of pediatric lumbosacral spondylolisthesis has been proposed recently. In this classification involving 8 distinct types of spondylolisthesis, the patient is classified according to: 1) the slip grade (low- vs. high-grade), 2) the degree of dysplasia (low- vs. high-dysplastic), and 3) the sagittal spinopelvic balance. The objective of this preliminary study is to assess the reliability of the classification. Two observers classified on two separate occasions 40 subjects with lumbosacral spondylolisthesis, based on standing postero-anterior and lateral radiographs of the spine and pelvis. No direct measurements on the radiographs were performed. All 8 types of spondylolisthesis were identified by the observers. Intra-observer agreement for the first and second observers was respectively 92.5% and 87.5%, while inter-observer agreement was 75.0%. Thirty-nine of 40 subjects had agreement among both observers according to the slip grade. Within these 39 subjects, observers disagreed for 8 subjects with respect to the degree of dysplasia and for only one subject with respect to the spinopelvic balance. The proposed classification could be used to better evaluate and compare available surgical techniques, and to develop a treatment algorithm for spondylolisthesis. This new classification results in good intra- and inter-observer agreement. Further studies with observers not involved in the design of the classification are however needed in order to confirm the relevance of the classification.

Humans↗

PubMed2002

Prospective validation of an algorithm with systematic sextant biopsy to predict pelvic lymph node metastasis in patients with clinically localized prostatic carcinoma.

PURPOSE: We prospectively validate an algorithm to predict pelvic lymph node metastasis in patients with clinically localized prostatic carcinoma. MATERIAL AND METHODS: A total of 293 patients with prostatic cancer were identified before pelvic lymph node dissection according to an algorithm developed with the classification and regression tree analysis as high-greater than 3 sextant biopsies containing any Gleason grade 4 or 5 cancer, intermediate-at least 1 biopsy dominated by Gleason grade 4 or 5 cancer but not high risk and low risk-all other patients. Observed and predicted frequencies of pelvic lymph node metastasis were compared. RESULTS: The observed frequencies of lymph node metastasis were remarkably similar to the predicted frequencies, including 2.8% versus 2.2% in 85.7% of patients in the low risk group, 16.7% versus 19.4% in 10.2% intermediate and 41.7% versus 45.5% in 4.1% high, respectively. If patients in the low risk group were considered to have node negative disease the specificity and negative predictive value of the algorithm were 88.4% and 97.2%, respectively. CONCLUSIONS: Our algorithm is valid as a simple and accurate tool for the prediction of pelvic lymph node metastasis in patients with clinically localized prostatic cancer. Those 85.7% of patients classified by the algorithm to have a low risk of lymphatic spread should not undergo pelvic lymph node dissection before definitive local treatment.

Algorithms↗

PubMed1997

Reinvestigation of a genetic-based classifier system: the effectiveness of recombination.

An empirical study for the effectiveness of recombination in a genetic-based classifier system applied to the field of ion chromatography is presented. From a comparison of the classifier system with and without crossover it followed that recombination was unable to make a significant contribution to the classification results. Despite this ineffectiveness, the genetic algorithm was a legitimate choice for solving the present classification problem because its population-based properties were of greater importance than the issue of whether or not recombination significantly added to the performance. These findings prompt the testing of other classifier systems as well, in order to reveal the extent to which the presented results can be classified.

Algorithms↗

PubMedDate not supplied

[It is normal for classification approaches to be diverse].

It is asserted that the postmodern concept of science, unlike the classical ideal, presumes necessary existence of various classification approaches (schools) in taxonomy, each corresponding to a particular aspect of consideration of the "taxic reality". They are set up by diversity of initial epistemological and ontological backgrounds which fix in a certain way a) fragments of that reality allowable for investigation, and b) allowable methods of exploration of the fragments being fixed. It makes it possible to define a taxonomic school as a unity of the above backgrounds together with consideration aspect delimited by them. Two extreme positions of these backgrounds could be recognized in recent taxonomic thought. One of them follows the scholastic tradition of elaboration of a formal and, hence, universal classificatory method ("new typology", numerical phenetics, pattern cladistics). Another one asserts dependence of classificatory approach on the judgment of the nature of taxic reality (natural philosophy, evolutionary schools of taxonomy). Some arguments are put forward in favor of significant impact of evolutionary thinking onto the theory of modern taxonomy. This impact is manifested by the correspondence principle which makes classificatory algorithms (and hence resulting classifications) depending onto initial assumptions about causes of taxic diversity. It is asserted that criteria of "quality" of both classifications proper and classificatory methods can be correctly formulated within the framework of a particular consideration aspect only. For any group of organisms, several particular classifications are rightful to exist, each corresponding to a particular consideration aspect. These classifications could not be arranged along the "better-worse" scale, as they reflect different fragments of the taxic reality. Their mutual interpretation depends on degree of compatibility of background assumptions and of the tasks being resolved. Extensionally, classifications are compatible as much as they coincide by context and hierarchical structure of included taxa. Intentionally, typological classifications are compatible if included taxa are comparable by their diagnoses, while phylogenetic classifications are compatible if the included taxa are ascribed monophyletic status. A brief consideration is given to the "new phylogenetics" (= "genophyletics") as to a classificatory approach aimed at elaboration of parsimonious phylogenetic hypotheses based on molecular biology data and employing numerical methods of cladistic analysis. This approach is shown to borrows some phenetic ideas and revives scholastic principle of unified classificatory basis. It is supposed that, in a time, biological classification would get escaping from plethora of positivistic ideas (including those being developed by nowaday cladistics) and would assimilate (revive) more actively holistic worldview.

Biodiversity↗

PubMed1999

Genomic signature: characterization and classification of species assessed by chaos game representation of sequences.

We explored DNA structures of genomes by means of a new tool derived from the "chaotic dynamical systems" theory (the so-called chaos game representation [CGR]), which allows the depiction of frequencies of oligonucleotides in the form of images. Using CGR, we observe that subsequences of a genome exhibit the main characteristics of the whole genome, attesting to the validity of the genomic signature concept. Base concentrations, stretches (runs of complementary bases or purines/pyrimidines), and patches (over- or underexpressed words of various lengths) are the main factors explaining the variability observed among sequences. The distance between images may be considered a measure of phylogenetic proximity. Eukaryotes and prokaryotes can be identified merely on the basis of their DNA structures.

Algorithms↗

PubMed2005

EDGE: a centralized resource for the comparison, analysis, and distribution of toxicogenomic information.

Transcriptional profiling via microarrays holds great promise for toxicant classification and hazard prediction. Unfortunately, the use of different microarray platforms, protocols, and informatics often hinders the meaningful comparison of transcriptional profiling data across laboratories. One solution to this problem is to provide a low-cost and centralized resource that enables researchers to share toxicogenomic data that has been generated on a common platform. In an effort to create such a resource, we developed a standardized set of microarray reagents and reproducible protocols to simplify the analysis of liver gene expression in the mouse model. This resource, referred to as EDGE, was then used to generate a training set of 117 publicly accessible transcriptional profiles that can be accessed at http://edge.oncology.wisc.edu/. The Web-accessible database was also linked to an informatics suite that allows on-line clustering and K-means analyses as well as Boolean and sequence-based searches of the data. We propose that EDGE can serve as a prototype resource for the sharing of toxicogenomics information and be used to develop algorithms for efficient chemical classification and hazard prediction.

Animals↗

PubMed2003

DNA microarray data and contextual analysis of correlation graphs.

BACKGROUND: DNA microarrays are used to produce large sets of expression measurements from which specific biological information is sought. Their analysis requires efficient and reliable algorithms for dimensional reduction, classification and annotation. RESULTS: We study networks of co-expressed genes obtained from DNA microarray experiments. The mathematical concept of curvature on graphs is used to group genes or samples into clusters to which relevant gene or sample annotations are automatically assigned. Application to publicly available yeast and human lymphoma data demonstrates the reliability of the method in spite of its simplicity, especially with respect to the small number of parameters involved. CONCLUSIONS: We provide a method for automatically determining relevant gene clusters among the many genes monitored with microarrays. The automatic annotations and the graphical interface improve the readability of the data. A C++ implementation, called Trixy, is available from http://tagc.univ-mrs.fr/bioinformatics/trixy.html.

Algorithms↗

PubMed2005

Gender and suicidality prediction in epilepsy.

The current study was carried out to compare the accuracy of suicidal risk prediction in patients with epilepsy for each gender separately and as a group (males+females). The discriminant function analysis was performed to create an algorithm for suicidal risk classification for males and females separately and as a group. The main characteristics of epilepsy, such as type and mean frequency of seizures and kind and mean daily dose of AEDs were used as independent variables. The accuracy of suicidality prediction is higher for each gender than for the joint group. There was a statistically significant difference between the joint group and male epileptic patients and a trend toward more precise suicidality prediction among men than women, although the discrepancies were not significant. It was concluded that gender is an important factor in suicidality prediction, and more precise prediction of suicidal risk is possible for male epileptic patients because of the better differentiation of suicide grades in males than in females.

Algorithms↗

PubMed2004

4SCOPmap: automated assignment of protein structures to evolutionary superfamilies.

BACKGROUND: Inference of remote homology between proteins is very challenging and remains a prerogative of an expert. Thus a significant drawback to the use of evolutionary-based protein structure classifications is the difficulty in assigning new proteins to unique positions in the classification scheme with automatic methods. To address this issue, we have developed an algorithm to map protein domains to an existing structural classification scheme and have applied it to the SCOP database. RESULTS: The general strategy employed by this algorithm is to combine the results of several existing sequence and structure comparison tools applied to a query protein of known structure in order to find the homologs already classified in SCOP database and thus determine classification assignments. The algorithm is able to map domains within newly solved structures to the appropriate SCOP superfamily level with approximately 95% accuracy. Examples of correctly mapped remote homologs are discussed. The algorithm is also capable of identifying potential evolutionary relationships not specified in the SCOP database, thus helping to make it better. The strategy of the mapping algorithm is not limited to SCOP and can be applied to any other evolutionary-based classification scheme as well. SCOPmap is available for download. CONCLUSION: The SCOPmap program is useful for assigning domains in newly solved structures to appropriate superfamilies and for identifying evolutionary links between different superfamilies.

Algorithms↗

PubMed1996

Merging Back-propagation and Hebbian Learning Rules for Robust Classifications.

By imposing saturation requirements on hidden-layer neural activations, a new learning algorithm is developed to improve robustness on classification performance of a multi-layer Perceptron. Derivatives of the sigmoid functions at hidden-layers are added to the standard output error with relative significance factors, and the total error is minimized by the steepest-descent method. The additional gradient-descent terms become Hebbian, and this new algorithm merges two popular learning algorithms, i.e., error back-propagation and Hebbian learning rules. Only slight modifications are needed for the standard back-propagation algorithm, and additional computational requirements are negligible. This saturation requirement effectively reduces output sensitivity to the input, which results in improved robustness and better generalization for classifier networks. Also distributed representations at hidden-layers are successfully suppressed to accomplish efficient utilization of hidden neurons. Computer simulations demonstrates much faster learning convergence as well as improved robustness for classifications and hetero-associations of binary patterns. Copyright 1996 Elsevier Science Ltd

Journal Article↗

PubMed1998

Body-surface map models for early diagnosis of acute myocardial infarction.

The standard 12-lead ECG is only 50% sensitive for the detection of acute myocardial infarction (AMI). The majority of leads for optimal classification of AMI probably lie outside the area covered by the 6 precordial leads. Thus, body-surface mapping (BSM) may be more helpful, as a larger thoracic area is sampled. We recorded 64-lead anterior BSMs in 635 patients with chest pain suggestive of AMI and abnormal electrocardiograms (ECGs), and 125 controls without chest pain. Of the 635 patients, 325 had AMI according to World Health Organization (WHO) criteria (203 presenting with ST segment elevation, and 122 with nondiagnostic ECG), and 310 had an "abnormal ECG but not AMI." QRS and ST-T isointegrals and variables describing map shape were derived. Subjects were randomly allocated to a training set (63 controls, 321 patients) and a validation set (62 controls, 314 patients). Multiple logistic regression was used in the training set to identify which variables gave best discrimination between groups. A model with these variables was then tested prospectively in the validation set. In stage 1 (all subjects), controls were compared with patients. In the training set, a model containing 21 variables classified 58/63 controls (specificity 92%) and 316/321 patients (sensitivity 98%). In the validation set, the model classified 48/62 controls (specificity 77.4%) and 302/314 patients (sensitivity 96%). In stage 2 (studying patients only), patients with AMI were compared with patients who had an abnormal ECG-not AMI. In the training set, a model containing 28 variables classified 132/165 patients (sensitivity 80%) with AMI and 134/156 patients (specificity 86%) with an abnormal ECG-not AMI. In the validation set, the model classified 123/160 patients (sensitivity 77%) with AMI and 131/154 patients (specificity 85%) with an abnormal ECG-not AMI. Combining results of both stages in a two-step algorithm gave an overall classification in the training set of controls 92%, abnormal ECG-not AMI 84%, AMI 80%, and in the validation set of controls 77%, abnormal ECG-not AMI 82%, AMI 74%. Thus, in conclusion, when compared with the 12-lead ECG, BSM models results in higher sensitivity and specificity for detection of AMI, particularly in patients presenting with chest pain and nondiagnostic ECG changes. The use of BSM models in such patients, may lead to the earlier detection of AMI and appropriate administration of fibrinolytic therapy and/or anti-platelet agents.

Body Surface Potential Mapping↗

PubMed1997

[Coronary perfusion: a classification based on the type and relative extension of coronary irrigation (II). An angiographic algorithm].

INTRODUCTION AND OBJECTIVES: An angiographic algorithm of clinical utility, applicable to conventional coronariography, is proposed to establish different patterns of coronary distribution depending on the characteristics of the myocardial perfusion, considering the starting point as the segmentary classification of the arterial irrigation of the left ventricle. METHODS: To validate this system of classification, 30 hearts coming from necropsy were studied, through anatomical and angiographical analysis. The average age of the population studied was of 69.8 +/- 14.6 years. The range was between 26 and 91 years. To study them, the hearts were unrolled and after a coronariography and a dissection of the coronary arterial tree, the identification of the perfusion mode--exclusive or shared--of every left ventricle segment was done. Then an algorithm based on the type of division of the left main branch, and on the type of perfusion of the left ventricle inferobasal segment was applied to the angiographic frames. There was statistical analysis of the data obtained in the anatomic and angiographic studies. To verify the applicability of the algorithm, it was employed to successive series of 100 coronariographies in vivo, and these were then compared to the results obtained with the necropsy series. RESULTS: The statistical comparison between the percentages of the classification obtained from both analyses of the necropsy series showed no significant differences. The statistical comparison of the percentages of the classification obtained between the in vivo and post-mortem analyses did not show any significant difference either. CONCLUSIONS: The angiographical algorithm developed allows to classify the myocardial perfusion of the left ventricle, by the conventional coronary arteriography, in three groups of clinical interest. The classification is based on the predominance of the left ventricular segments exclusively irrigated by: the anterior interventricular artery (type I), the circumflex artery (type II), or a balance between both arteries (type III). The angiographic projections in left anterior oblique with caudal angulation and right anterior oblique are important for its application. The classification of the ventricular perfusion established with the developed algorithm can be validated as being equivalent to the one obtained through the anatomical series.

Adult↗

PubMed1986

[Prognostication of individual radiation sensitivity of animals, based on the use of a mathematical method of object classification].

A statistical method of object classification based on the use of linear discriminant analysis, was employed to solve the problem of prediction of the outcome of acute radiation disease. In experiments on 23 monkeys irradiated at a dose of 5.3 Gy 14 animals had survived and 9 perished. The initial state and early response (in 5 h) to gamma-beam radiation were assessed. The total number of the analyzed parameters was 16. According to the experimental results an integral index--discriminator--was computed for each object. Values for the survivors ranged within 136.1-151.3, and for 9 more radiosensitive animals the values of discriminators turned out to be higher (151.9-168.7). The procedure of prognosis was the following. 23 monkeys were randomly divided into 2 groups (using the table of random numbers). The 1st group (a study sample) included 13 animals, of them 8 survived and 5 perished. The 2nd group (a control predictor sample) included 10 animals, of them 6 survived and 4 perished by the end of the experiment. The authors described 3 variants of a random selection of the animals both for the 1st and 2nd groups. The results of prognosis were compared with the experimental results. Erroneous prognosis for each variant was 10, 0 and 20%, respectively.

Algorithms↗