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 1,081 records · Page 60Linked to original sources

PubMed1983

Digital analysis of RR intervals for identification of cardiac arrhythmias.

This paper describes a digital computer program for the automated detection and classification of persistent cardiac rhythms. The algorithm is based on dynamic mathematical models and state estimation. The method is based on the use of RR interval data and the development of simple models that describe the sequential behavior of such intervals characteristic of different persistent rhythms. The method involves the calculation of sufficient statistics for the data. The method was tested on the actual rhythm data; it provides a rapid and accurate identification of persistent cardiac rhythms.

Arrhythmias, Cardiac↗

PubMed1990

Electrocardiographic tall R waves in the right precordial leads. Comparison of recently proposed ECG and VCG criteria for distinguishing posterolateral myocardial infarction from prominent anterior forces in normal subjects.

Electrocardiographic tall R waves in the right precordial leads may be present in patients with posterior myocardial infarction, right ventricular hypertrophy, various conduction disturbances, and some forms of cardiomyopathy and in clinically otherwise normal subjects with prominent anterior electromotive forces. Clinical uncertainty most often arises in distinguishing possible prior posterolateral myocardial infarction (PMI) from the unusual normal variant (PAF). The ECGs and VCGs of 15 subjects with posterolateral infarction were compared with tracings from 12 subjects with no evidence of cardiac disease, all individuals demonstrating tall R waves (R/S greater than 1.0 in V1 and/or V2) in the right precordial leads on surface ECG. By standard ECG, the infarction group was characterized by taller T waves in leads V1 and V2, shorter T waves in V6, greater T2-T6 index, and a more negative two variable function as described by Nestico. By VCG, the infarction group was characterized by a more anteriorly oriented T loop, more leftward maximal frontal plane QRS vector and a lower calculated -45 degrees/ab, as described by Suzuki. An algorithm was proposed that permitted proper classification (PAF vs. PMI) based on ECG criteria in 75% of subjects with 90% accuracy. This compared favorably with performance of the Frank vectorcardiogram, including using more recently proposed criteria. Routine use of the VCG, therefore, in this clinical setting may no longer be justified.

Diagnosis, Differential↗

PubMed1988

Unsupervised waveform classification for multi-neuron recordings: a real-time, software-based system. I. Algorithms and implementation.

We describe a new, mostly software-based device for the sorting of waveforms in an extracellular multi-neuron recording situation. The sorting algorithm is largely unattended, and, after an initial 'learning' process, works in real time. Shape comparisons are based on up to 8 time points in the waveform; these points (the reduced feature set) are chosen automatically by analyzing the current incoming data stream. A feasibility version has been implemented on a LSI-11/2 system, using FORTRAN for set-up calculations and assembler for the real-time operations. Detailed comparisons with performance of other sorting devices are presented in the companion paper.

Algorithms↗

PubMed2006

Anatomical sketch understanding: recognizing explicit and implicit structure.

OBJECTIVE: Sketching is ubiquitous in medicine. Physicians commonly use sketches as part of their note taking in patient records and to help convey diagnoses and treatments to patients. Medical students frequently use sketches to help them think through clinical problems in individual and group problem solving. Applications ranging from automated patient records to medical education software could benefit greatly from the richer and more natural interfaces that would be enabled by the ability to understand sketches. In this paper we take the first steps toward developing a system that can understand anatomical sketches. METHODS: Understanding an anatomical sketch requires the ability to recognize what anatomical structure has been sketched and from what view (e.g. parietal view of the brain), as well as to identify the anatomical parts and their locations in the sketch (e.g. parts of the brain), even if they have not been explicitly drawn. We present novel algorithms for sketch recognition and for part identification. We evaluate the accuracy of the recognition algorithm on sketches obtained from medical students. We evaluate the part identification algorithm by comparing its results to the judgment of an experienced physician. RESULTS: The sketch recognition algorithm achieves a recognition accuracy of 75.5%, far above the baseline random classification accuracy of 6.7%. Comparison of the results of the part identification algorithm with the judgment of an experienced physician shows close agreement in terms of location, orientation, size, and shape of the identified parts. CONCLUSIONS: The performance of our prototype in terms of accuracy and running time provides strong evidence that development of robust sketch understanding systems for medical domains is an attainable goal. Further work needs to be done to extend the approach to sketches containing multiple and partial anatomical structures, as well as to be able to interpret sketch annotations.

Algorithms↗

PubMedDate not supplied

Training neural networks with heterogeneous data.

Data pruning and ordered training are two methods and the results of a small theory that attempts to formalize neural network training with heterogeneous data. Data pruning is a simple process that attempts to remove noisy data. Ordered training is a more complex method that partitions the data into a number of categories and assigns training times to those assuming that data size and training time have a polynomial relation. Both methods derive from a set of premises that form the 'axiomatic' basis of our theory. Both methods have been applied to a time-delay neural network-which is one of the main learners in Microsoft's Tablet PC handwriting recognition system. Their effect is presented in this paper along with a rough estimate of their effect on the overall multi-learner system. The handwriting data and the chosen language are Italian.

Algorithms↗

PubMedDate not supplied

Incremental learning of feature space and classifier for face recognition.

We have proposed a new approach to pattern recognition in which not only a classifier but also a feature space of input variables is learned incrementally. In this paper, an extended version of Incremental Principal Component Analysis (IPCA) and Resource Allocating Network with Long-Term Memory (RAN-LTM) are effectively combined to implement this idea. Since IPCA updates a feature space incrementally by rotating the eigen-axes and increasing the dimensions, the inputs of a neural classifier must also change in their values and the number of input variables. To solve this problem, we derive an approximation of the update formula for memory items, which correspond to representative training samples stored in the long-term memory of RAN-LTM. With these memory items, RAN-LTM is efficiently reconstructed and retrained to adapt to the evolution of the feature space. This function is incorporated into our face recognition system. In the experiments, the proposed incremental learning model is evaluated over a self-compiled video clip of 24 persons. The experimental results show that the incremental learning of a feature space is very effective to enhance the generalization performance of a neural classifier in a realistic face recognition task.

Algorithms↗

PubMedDate not supplied

Handling missing values in support vector machine classifiers.

This paper discusses the task of learning a classifier from observed data containing missing values amongst the inputs which are missing completely at random. A non-parametric perspective is adopted by defining a modified risk taking into account the uncertainty of the predicted outputs when missing values are involved. It is shown that this approach generalizes the approach of mean imputation in the linear case and the resulting kernel machine reduces to the standard Support Vector Machine (SVM) when no input values are missing. Furthermore, the method is extended to the multivariate case of fitting additive models using componentwise kernel machines, and an efficient implementation is based on the Least Squares Support Vector Machine (LS-SVM) classifier formulation.

Algorithms↗

PubMed2004

CRUISE: cortical reconstruction using implicit surface evolution.

Segmentation and representation of the human cerebral cortex from magnetic resonance (MR) images play an important role in neuroscience and medicine. A successful segmentation method must be robust to various imaging artifacts and produce anatomically meaningful and consistent cortical representations. A method for the automatic reconstruction of the inner, central, and outer surfaces of the cerebral cortex from T1-weighted MR brain images is presented. The method combines a fuzzy tissue classification method, an efficient topology correction algorithm, and a topology-preserving geometric deformable surface model (TGDM). The algorithm is fast and numerically stable, and yields accurate brain surface reconstructions that are guaranteed to be topologically correct and free from self-intersections. Validation results on real MR data are presented to demonstrate the performance of the method.

Algorithms↗

PubMed2003

Detection of single unit activity from the rat vagus using cluster analysis of principal components.

In vivo recordings from subdiaphragmatic vagal afferent nerves generally lack the resolution to distinguish single unit activity. Several methods for data acquisition and analysis were combined to produce a high degree of reliability in recording electrophysiological signals from gastrointestinal and hepatic afferent fibers in the rat. Recordings with low noise were achieved by paralysis of the respiratory muscles and by pinning the nerve to a recording platform. Single unit activity was isolated using principal component (PC) analysis and cluster cutting of data in multi-dimensional space (1-3 PCs). Cluster assignments were determined by a semi-automated approach using the k-means algorithm. The accuracy of single unit classification was assessed by checking inter-spike intervals (ISIs) to determine the length of the refractory period, and by cross-correlation analysis to assess whether single units were mistakenly split into more than one cluster. These analyses produced up to four isolated single units from each nerve filament (a bundle of nerve fibers), and typically it was possible to further increase yield by recording from several nerve filaments simultaneously using an array of electrodes.

Action Potentials↗

PubMed2003

Narrative electronic patient records as source of discharge diagnoses.

OBJECTIVE: To evaluate the consistency of diagnostic data extracted from narrative electronic patient record (EPR) notes compared with the data from a patient administrative system (PAS). To assess potential benefit of using EPR notes as source of diagnosis data and as basis for case identification. DESIGN: Construction of a computer algorithm to extract ICD-9 codes from narrative EPR notes. Assessment of consistency and reliability of the diagnostic codes retrieved from EPR notes and PAS. Estimation of efficiency of case identification based on data from PAS and EPR. RESULTS: Diagnosis codes were retrieved from PAS with sensitivity of 0.989 and the positive predictive value (PPV) was 0.993. Codes were retrieved from EPR with sensitivity of 0.908 and PPV of 0.990. Combining these two sources increased sensitivity to 0.999. CONCLUSION: Discharge diagnoses were easily extracted from narrative EPR notes by automatic methods. Information extracted from record notes was not significantly different from the corresponding data in PAS, but EPR was incomplete as compared with PAS. Utilizing data extracted from EPR improved case identification significantly.

Algorithms↗

PubMed1999

Rule-based programming paradigm: a formal basis for biological, chemical and physical computation.

A rule-based programming paradigm is described as a formal basis for biological, chemical and physical computations. In this paradigm, the computations are interpreted as the outcome arising out of interaction of elements in an object space. The interactions can create new elements (or same elements with modified attributes) or annihilate old elements according to specific rules. Since the interaction rules are inherently parallel, any number of actions can be performed cooperatively or competitively among the subsets of elements, so that the elements evolve toward an equilibrium or unstable or chaotic state. Such an evolution may retain certain invariant properties of the attributes of the elements. The object space resembles Gibbsian ensemble that corresponds to a distribution of points in the space of positions and momenta (called phase space). It permits the introduction of probabilities in rule applications. As each element of the ensemble changes over time, its phase point is carried into a new phase point. The evolution of this probability cloud in phase space corresponds to a distributed probabilistic computation. Thus, this paradigm can handle tor deterministic exact computation when the initial conditions are exactly specified and the trajectory of evolution is deterministic. Also, it can handle probabilistic mode of computation if we want to derive macroscopic or bulk properties of matter. We also explain how to support this rule-based paradigm using relational-database like query processing and transactions.

Algorithms↗

PubMed2001

Analysis of codon usage diversity of bacterial genes with a self-organizing map (SOM): characterization of horizontally transferred genes with emphasis on the E. coli O157 genome.

With increases in the amounts of available DNA sequence data, it has become increasingly important to develop tools for comprehensive systematic analysis and comparison of species-specific characteristics of protein-coding sequences for a wide variety of genomes. In the present study, we used a novel neural-network algorithm, a self-organizing map (SOM), to efficiently and comprehensively analyze codon usage in approximately 60,000 genes from 29 bacterial species simultaneously. This SOM makes it possible to cluster and visualize genes of individual species separately at a much higher resolution than can be obtained with principal component analysis. The organization of the SOM can be explained by the genome G+C% and tRNA compositions of the individual species. We used SOM to examine codon usage heterogeneity in the E. coli O157 genome, which contains 'O157-unique segments' (O-islands), and showed that SOM is a powerful tool for characterization of horizontally transferred genes.

Algorithms↗

PubMedDate not supplied

An accelerated procedure for recursive feature ranking on microarray data.

We describe a new wrapper algorithm for fast feature ranking in classification problems. The Entropy-based Recursive Feature Elimination (E-RFE) method eliminates chunks of uninteresting features according to the entropy of the weights distribution of a SVM classifier. With specific regard to DNA microarray datasets, the method is designed to support computationally intensive model selection in classification problems in which the number of features is much larger than the number of samples. We test E-RFE on synthetic and real data sets, comparing it with other SVM-based methods. The speed-up obtained with E-RFE supports predictive modeling on high dimensional microarray data.

Oligonucleotide Array Sequence Analysis↗

PubMed1997

Response to chemotherapy has predictive value for further survival of patients with advanced non-small cell lung cancer: 10 years experience of the European Lung Cancer Working Party.

The aim of this study was the assessment of the predictive value for survival of an antitumoral response to three courses of chemotherapy in association with various pretreatment characteristics in patients with non-resectable non-small cell lung cancer treated by cisplatin- (or carboplatin)-based combination regimens. Patients considered for this study were eligible patients with advanced non-small cell lung cancer registered in one of the seven trials conducted by the European Lung Cancer Working Party from December 1980 to August 1991. All these trials tested chemotherapy regimens with platinum derivatives (cisplatin and/or carboplatin). In this population of 1052 eligible patients, 752 were assessed in this analysis. Data were prospectively collected on 23 pretherapeutic variables and objective response after three chemotherapy cycles. The predictive value of response to chemotherapy on survival (measured from the time of response assessment i.e. 12 weeks after registration in the trial) was studied by univariate analysis as well as by multivariate methods (adjustment of the impact of several covariates simultaneously on the dependent variable) with adjustment for the pretreatment prognostic variables. After three cycles of chemotherapy, the global estimated median survival time was 24 weeks with a 95% confidence interval of 22-25 weeks. By univariate analysis, we identified an objective response to chemotherapy as a highly significant discriminant marker (P < 0.0001) for further survival with estimated median survival times of 41 weeks (95% CI: 38-46) and 19 weeks (95% CI: 17-20), respectively, for the responding and non-responding patients. In a Cox regression model fitted to the data using a forward stepwise procedure, this variable was the first selected explanatory variable. Its effect was adjusted by the introduction in the model of initial disease extent, Karnofsky performance status, serum calcium level and white blood cell count. These results were consistent with those obtained by application of recursive partitioning and amalgamation algorithms (RECPAM) which led to a classification of the patients into three homogeneous subgroups. Our results, using a classical Cox regression model consistent with those highlighted by application of a RECPAM analysis, found an objective response to chemotherapy to be a predominant predictive factor for further survival, although it did not allow any conclusion about a causal relationship. The RECPAM results led to a classification of the patients into three subgroups which needs to be validated in other series.

Adult↗

PubMed2004

An empirical test of the treatment of indels during optimization alignment based on the phylogeny of the genus Secale (Poaceae).

The ability of the program POY, implementing optimization alignment, to deal with major indels is explored and discussed in connection with a phylogenetic analysis of the genus Secale based on partial Adh1 sequences. The Adh1 sequences used span exon 2-4. Nearly all variation is found in intron 2 and intron 3, which form the basis for the phylogenetic analyses. Both in some ingroup and outgroup taxa intron 3 has a major duplication. Previous phylogenetic analyses have repeatedly confirmed monophyly of both Secale and Hordeum, the latter being part of the outgroup. However, optimization alignment only recovers both genera as monophyletic when knowledge of the duplication is incorporated in the analysis. The phylogenetic relationships within Secale are not clearly resolved. Subspecific taxa of Secale strictum have identical sequences and they are confined to a monophyletic group. However, the two subspecific taxa of Secale cereale do not form a monophyletic group, and the position of Secale sylvestre is uncertain.

Algorithms↗

PubMed1997

Segmentation of the visible human for high-quality volume-based visualization.

This article describes a combination of interactive classification and super-sampling visualization algorithms that greatly enhances the realism of 3-D reconstructions of the Visible Human data sets. Objects are classified on the basis of ellipsoidal regions in RGB space. The ellipsoids are used for super-sampling in the visualization process.

Abdomen↗

PubMed2003

Multivariate analysis and classification of the chemical quality of 7-aminocephalosporanic acid using near-infrared reflectance spectroscopy.

The capability of near-infrared (NIR) spectroscopy in comparison to conventional chemical testing to control the chemical quality of a pharmaceutical intermediate has been investigated. Multivariate projection methods including principal component analysis, partial least-squares discriminant analysis and soft independent modeling of class analogy have been evaluated. 7-Aminocephalosporanic acid has been chosen as an example providing a large variation of quality due to its relative chemical instability. Three sets of production lots have been selected to study the extent of quality information extractable from NIR spectra. The first set of 91 lots covers a very broad range of chemical quality assessed by 8 parameters with a partially extended characterization by physical properties. The general congruence of spectral, chemical, and physical information has been investigated. The second set of 110 lots covers a very narrow range of chemical quality assessed by 11 parameters. With extended quality information, the intrinsic selectivity within the spectral data structure has been studied. The third set of 228 lots characterized by 8 parameters is a selection out of more than 1000 lots over a production period of two years. The ruggedness of the multivariate approach has been confirmed by a cross validation of the classification test.

Algorithms↗

PubMed2007

Comparative urine analysis by liquid chromatography-mass spectrometry and multivariate statistics: method development, evaluation, and application to proteinuria.

We describe a platform for the comparative profiling of urine using reversed-phase liquid chromatography-mass spectrometry (LC-MS) and multivariate statistical data analysis. Urinary compounds were separated by gradient elution and subsequently detected by electrospray Ion-Trap MS. The lower limit of detection (5.7-21 nmol/L), within-day (2.9-19%) and between-day (4.8-19%) analytical variation of peak areas, linearity (R2: 0.918-0.999), and standard deviation for retention time (<0.52 min) of the method were assessed by means of addition of seven 3-8 amino acid peptides (0-500 nmol/L). Relating the amount of injected urine to the area under the curve (AUC) of the chromatographic trace at 214 nm better reduced the coefficient of variation (CV) of the AUC of the total ion chromatogram (CV = 10.1%) than relating it to creatinine (CV = 38.4%). LC-MS data were processed, and the common peak matrix was analyzed by principal component analysis (PCA) after supervised classification by the nearest shrunken centroid algorithm. The feasibility of the method to discriminate urine samples of differing compositions was evaluated by (i) addition of seven peptides at nanomolar concentrations to blank urine samples of different origin and (ii) a study of urine from kidney patients with and without proteinuria. (i) The added peptides were ranked as highly discriminatory peaks despite significant biological variation. (ii) Ninety-two peaks were selected best discriminating proteinuric from nonproteinuric samples, of which 6 were more intense in the majority of the proteinuric samples. Two of these 6 peaks were identified as albumin-derived peptides, which is in accordance with the early rise of albumin during glomerular proteinuria. Interestingly, other albumin-derived peptides were nondiscriminatory indicating preferential proteolysis at some cleavage sites.

Algorithms↗