Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “CLASSIFICATION”

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 955 records · Page 53Linked to original sources
PubMed1998

Feature selection for genetic sequence classification.

MOTIVATION: Most of the existing methods for genetic sequence classification are based on a computer search for homologies in nucleotide or amino acid sequences. The standard sequence alignment programs scale very poorly as the number of sequences increases or the degree of sequence identity is <30%. Some new computationally inexpensive methods based on nucleotide or amino acid compositional analysis have been proposed, but prediction results are still unsatisfactory and depend on the features chosen to represent the sequences. RESULTS: In this paper, a feature selection method based on the Gamma (or near-neighbour) test is proposed. If there is a continuous or smooth map from feature space to the classification target values, the Gamma test gives an estimate for the mean-squared error of the classification, despite the fact that one has no a priori knowledge of the smooth mapping. We can search a large space of possible feature combinations for a combination which gives a smallest estimated mean-squared error using a genetic algorithm. The method was used for feature selection and classification of the large subunits of rRNA according to RDP (Ribosomal Database Project) phylogenetic classes. The sequences were represented by dinucleotide frequency distribution. The nearest-neighbour criterion has been used to estimate the predictive accuracy of the classification based on the selected features. For examples discussed, we found that the classification according to the first nearest neighbour is correct for 80% of the test samples. If we consider the set of the 10 nearest neighbours, then 94% of the test samples are classified correctly. AVAILABILITY: The principal novel component of this method is the Gamma test and this can be downloaded compiled for Unix Sun 4, Windows 95 and MS-DOS from http://www.cs.cf.ac.uk/ec/ CONTACT: s.margetts@cs.cf.ac.uk

Algorithms↗
PubMed2002

Building an automated classification of DNA-binding protein domains.

Intensive growth in 3D structure data on DNA-protein complexes as reflected in the Protein Data Bank (PDB) demands new approaches to the annotation and characterization of these data and will lead to a new understanding of critical biological processes involving these data. These data and those from other protein structure classifications will become increasingly important for the modeling of complete proteomes. We propose a fully automated classification of DNA-binding protein domains based on existing 3D-structures from the PDB. The classification, by domain, relies on the Protein Domain Parser (PDP) and the Combinatorial Extension (CE) algorithm for structural alignment. The approach involves the analysis of 3D-interaction patterns in DNA-protein interfaces, assignment of structural domains interacting with DNA, clustering of domains based on structural similarity and DNA-interacting patterns. Comparison with existing resources on describing structural and functional classifications of DNA-binding proteins was used to validate and improve the approach proposed here. In the course of our study we defined a set of criteria and heuristics allowing us to automatically build a biologically meaningful classification and define classes of functionally related protein domains. It was shown that taking into consideration interactions between protein domains and DNA considerably improves the classification accuracy. Our approach provides a high-throughput and up-to-date annotation of DNA-binding protein families which can be found at http://spdc.sdsc.edu.

Artificial Intelligence↗
PubMed2004

Systematic benchmarking of microarray data classification: assessing the role of non-linearity and dimensionality reduction.

MOTIVATION: Microarrays are capable of determining the expression levels of thousands of genes simultaneously. In combination with classification methods, this technology can be useful to support clinical management decisions for individual patients, e.g. in oncology. The aim of this paper is to systematically benchmark the role of non-linear versus linear techniques and dimensionality reduction methods. RESULTS: A systematic benchmarking study is performed by comparing linear versions of standard classification and dimensionality reduction techniques with their non-linear versions based on non-linear kernel functions with a radial basis function (RBF) kernel. A total of 9 binary cancer classification problems, derived from 7 publicly available microarray datasets, and 20 randomizations of each problem are examined. CONCLUSIONS: Three main conclusions can be formulated based on the performances on independent test sets. (1) When performing classification with least squares support vector machines (LS-SVMs) (without dimensionality reduction), RBF kernels can be used without risking too much overfitting. The results obtained with well-tuned RBF kernels are never worse and sometimes even statistically significantly better compared to results obtained with a linear kernel in terms of test set receiver operating characteristic and test set accuracy performances. (2) Even for classification with linear classifiers like LS-SVM with linear kernel, using regularization is very important. (3) When performing kernel principal component analysis (kernel PCA) before classification, using an RBF kernel for kernel PCA tends to result in overfitting, especially when using supervised feature selection. It has been observed that an optimal selection of a large number of features is often an indication for overfitting. Kernel PCA with linear kernel gives better results.

Algorithms↗
PubMed2006

Effects of replacing the unreliable cDNA microarray measurements on the disease classification based on gene expression profiles and functional modules.

MOTIVATION: Microarrays datasets frequently contain a large number of missing values (MVs), which need to be estimated and replaced for subsequent data mining. The focus of the paper is to study the effects of different MV treatments for cDNA microarray data on disease classification analysis. RESULTS: By analyzing five datasets, we demonstrate that among three kinds of classifiers evaluated in this study, support vector machine (SVM) classifiers are robust to varied MV imputation methods [e.g. replacing MVs by zero, K nearest-neighbor (KNN) imputation algorithm, local least square imputation and Bayesian principal component analysis], while the classification and regression tree classifiers are sensitive in terms of classification accuracy. The KNNclassifiers built on differentially expressed genes (DEGs) are robust to the varied MV treatments, but the performances of the KNN classifiers based on all measured genes can be significantly deteriorated when imputing MVs for genes with larger missing rate (MR) (e.g. MR > 5%). Generally, while replacing MVs by zero performs relatively poor, the other imputation algorithms have little difference in affecting classification performances of the SVM or KNN classifiers. We further demonstrate the power and feasibility of our recently proposed functional expression profile (FEP) approach as means to handle microarray data with MVs. The FEPs, which are derived from the functional modules that are enriched with sets of DEGs and thus can be consistently identified under varied MV treatments, achieve precise disease classification with better biological interpretation. We conclude that the choice of MV treatments should be determined in context of the later approaches used for disease classification. The suggested exclusion criterion of ignoring the genes with larger MR (e.g. >5%), while justifiable for some classifiers such as KNN classifiers, might not be considered as a general rule for all classifiers.

Algorithms↗
PubMed1992

International primary care classifications: the effect of fifteen years of evolution.

To better understand the development of primary care classifications over the past 15 years, 10 primary care databases have been retrospectively analysed using the structure of the International Classification of Primary Care (ICPC) as the basis. All datasets were based on routine data collection using different classification systems by several family physicians during all encounters with their patients over considerable periods of time, in most cases one year. The prevalences or the rates of the available diagnostic--and reason for encounter--classes were distributed over four frequencies. With a few exceptions the distribution of diagnostic labels referring to common diseases is surprisingly similar. The use of ICPC however results in a quantum leap in the use of symptom and complaint diagnoses. Because of this shift primary care physicians now have available a classification with 400 diagnostic classes used with a prevalence of > or = 1/1000 patient-years or per 1000 visiting patients per year. The classification of reasons for encounter allows the physician to identify over 300 reasons for encounter used > or = 1/1000 patient years or per 1000 visiting patients per year. Family physicians have been successful in the development of new primary care classifications. Rag bag rubrics which are the result of the structure of ICPC are used relatively often and deserve more attention from primary care taxonomers.

Humans↗
PubMed2003

The use of occupation and industry classifications in general population studies.

Occupation and industry classifications are used in epidemiological studies to classify study subjects according to their job and subsequently to study risk by job, to infer social class indicators, or to infer exposure to specific agents through job-exposure matrices. However, documentation on methodological aspects concerning the use of occupation and industry classifications is sparse within epidemiology. This paper reviews the diverse applications of occupation and industry classifications in population-based epidemiological studies. The different classifications in use are discussed, and criteria are given for choosing a classification in an epidemiological study. Finally, the reliability of coding for occupation and industry is reviewed. A further standardization of the use of occupation and industry classifications in epidemiology is recommended, in order to facilitate future comparisons between studies and fully exploit their possibilities, especially when occupational exposures are to be inferred.

Databases, Factual↗
PubMed2004

SCOR: Structural Classification of RNA, version 2.0.

SCOR, the Structural Classification of RNA (http://scor.lbl.gov), is a database designed to provide a comprehensive perspective and understanding of RNA motif three-dimensional structure, function, tertiary interactions and their relationships. SCOR 2.0 represents a major expansion and introduces a new classification organization. The new version represents the classification as a Directed Acyclic Graph (DAG), which allows a classification node to have multiple parents, in contrast to the strictly hierarchical classification used in SCOR 1.2. SCOR 2.0 supports three types of query terms in the updated search engine: PDB or NDB identifier, nucleotide sequence and keyword. We also provide parseable XML files for all information. This new release contains 511 RNA entries from the PDB as of 15 May 2003. A total of 5880 secondary structural elements are classified: 2104 hairpin loops and 3776 internal loops. RNA motifs reported in the literature, such as 'Kink turn' and 'GNRA loops', are now incorporated into the structural classification along with definitions and descriptions.

Animals↗
PubMed2006

Topology-based cancer classification and related pathway mining using microarray data.

Cancer classification is the critical basis for patient-tailored therapy, while pathway analysis is a promising method to discover the underlying molecular mechanisms related to cancer development by using microarray data. However, linking the molecular classification and pathway analysis with gene network approach has not been discussed yet. In this study, we developed a novel framework based on cancer class-specific gene networks for classification and pathway analysis. This framework involves a novel gene network construction, named ordering network, which exhibits the power-law node-degree distribution as seen in correlation networks. The results obtained from five public cancer datasets showed that the gene networks with ordering relationship are better than those with correlation relationship in terms of accuracy and stability of the classification performance. Furthermore, we integrated the ordering networks, classification information and pathway database to develop the topology-based pathway analysis for identifying cancer class-specific pathways, which might be essential in the biological significance of cancer. Our results suggest that the topology-based classification technology can precisely distinguish cancer subclasses and the topology-based pathway analysis can characterize the correspondent biochemical pathways even if there are subtle, but consistent, changes in gene expression, which may provide new insights into the underlying molecular mechanisms of tumorigenesis.

Gene Expression Profiling↗
PubMed2000

Pathologists' agreement with experts and reproducibility of breast ductal carcinoma-in-situ classification schemes.

Several histologic classifications for breast ductal carcinoma in situ (DCIS) have been proposed. This study assessed the diagnostic agreement and reproducibility of three DCIS classifications (Holland [HL], modified Lagios [LA], and Van Nuys [VN]) by comparing the interpretations of pathologists without expertise in breast pathology with those of three breast pathology experts, each a proponent of one classification. Seven nonexpert pathologists in New Hampshire and three experts evaluated 40 slides of DCIS according to the three classifications. Twenty slides were reinterpreted by each nonexpert pathologist. Diagnostic accuracy (nonexperts compared with experts) and reproducibility were evaluated using inter- and intrarater techniques (kappa statistic). Final DCIS grade and nuclear grade were reported most accurately among nonexpert pathologists using HL (kappa = 0.53 and 0.49, respectively) compared with LA and VN (kappa = 0.29 and 0.35, respectively, for both classifications). An intermediate DCIS grade was assessed most accurately using HL and LA, and a high grade (group 3) was assessed most accurately using VN. Diagnostic reproducibility was highest using HL (kappa = 0.49). The VN interpretation of necrosis (present or absent) was reported more accurately than the LA criteria (extensive, focal, or absent; kappa = 0.59 and 0.45, respectively), but reproducibility of each was comparable (kappa = 0.48 and 0.46, respectively). Intrarater agreement was high overall. Comparing all three classifications, final DCIS grade was reported best using HL. Nuclear grade (cytodifferentiation) using HL and the presence or absence of necrosis were the criteria diagnosed most accurately and reproducibly. Establishing one internationally approved set of interpretive definitions, with acceptable accuracy and reproducibility among both pathologists with and without expertise in breast pathology interpretation, will assist researchers in evaluating treatment effectiveness and characterizing the natural history of DCIS breast lesions.

Breast Neoplasms↗
PubMed1994

Definition and classification of negative outcomes in solid organ transplantation. Application in liver transplantation.

OBJECTIVE: This study defined negative outcomes of solid organ transplantation, proposed a new classification of complications by severity, and applied the classification to evaluate the results of orthotopic liver transplantation (OLT). SUMMARY AND BACKGROUND DATA: The lack of uniform reporting of negative outcomes has made reports of transplantation procedures difficult to interpret and compare. In fact, only mortality is well reported; morbidity rates and severity of complications have been poorly described. METHODS: Based on previous definition and classification of complications for general surgery, a new classification for transplantation in four grades is proposed. Results including risk factors of the first 215 OLTs performed at the University of Toronto have been evaluated using the classification. RESULTS: All but two patients (99%) had at least one complication of any kind, 92% of patients surviving more than 3 months had grade 1 (minor) complications, 74% had grade 2 (life-threatening) complications, and 30% had grade 3 (residual disability or cancer) complications. Twenty-nine per cent of patients had grade 4 complications (retransplantation or death). The most common grade 1 complications were steroid responsive rejection (69% of patients) and infection that did not require antibiotics or invasive procedures (23%). Grade 2 complications primarily were infection requiring antibiotics or invasive procedures (64%), postoperative bleeding requiring > 3 units of packed red cells (35%), primary dysfunction (26%), and biliary disease treated with antibiotics or requiring invasive procedures (18%). The most frequent grade 3 complication was renal failure, which is defined as a permanent rise in serum creatinine levels > or = twice the pretransplantation values (11%). Grade 4 complications (retransplantation or death) mainly were infection (14%) and primary dysfunction (11%). Comparison between the first and last 50 OLTs of the series indicates a significant decrease in the mean number of grade 1 and 2 complications. This was partially a result of better medical status of patients at the time of transplantation. Using univariate and multivariate analyses of risk factors, the best predictor of grade 1 complications was donor obesity; for grade 2 complications, the best predictor was a donor liver rewarming time of > 90 minutes, and for grade 3 and 4 complications, the best predictor was the APACHE II scoring system and donor cardiac arrest. CONCLUSIONS: Standardized definitions and classifications of complications of transplantation will allow us to better evaluate and compare results of transplantation among centers and over time, and better compare effectiveness of new therapies. Orthotopic liver transplantation still is a procedure with high morbidity that requires careful analysis of risk factors to optimize selection of patients and organ sharing.

Adolescent↗
PubMed2003

Classification of gunshot injuries in civilians.

Gunshot injuries have become extremely prevalent among the United States civilian population because of increasing urban violence and the availability of handguns. However, the increasing prevalence of gunshot injuries in civilians dramatically contrasts with the paucity of scientific literature pertaining to the diagnosis, classification, and treatment of these injuries. The objective of the current study was to delineate the principal factors associated with gunshot injury severity in civilians, designate their importance in various injury patterns, and propose a new comprehensive classification system that may establish more uniform treatment approaches. The authors critically review existing gunshot injury classification systems with emphasis on the ballistic and clinical parameters that compose each system. The authors propose a new classification system based on five gunshot injury parameters: energy, vital structures involved, wound characteristics, fracture, and degree of contamination. This new classification scheme is applicable to all firearm injuries in civilians and assists with proper treatment selection. The proposed classification system is based on the authors' clinical experience in a Level I urban trauma center, and will require validation in a prospective clinical setting.

Fractures, Comminuted↗
PubMed2003

Gunshot femoral shaft fractures: is the current classification system reliable?

The reliability of the AO/Orthopaedic Trauma Association classification system has not been evaluated for diaphyseal fractures or fractures attributable to gunshot injuries. Therefore, the current authors assessed its reliability for diaphyseal femur fractures and investigated the effect of a gunshot mechanism of injury. Forty-seven diaphyseal femur fractures, 23 caused by gunshots and 24 caused by blunt trauma, were classified by four observers on two occasions. The interobserver and intraobserver reliability of each level of the AO/Orthopaedic Trauma Association classification was assessed with kappa statistics. Determination of fracture type had substantial interobserver and intraobserver reliability for gunshot and blunt injuries. Reliability decreased at the subsequent levels of the classification. Fractures caused by gunshots compared with those caused by blunt trauma were characterized by significantly lower interobserver agreement on fracture group (k = 0.26 versus 0.45) and subgroup (k = 0.21 versus 0.38). The AO/Orthopaedic Trauma Association classification system has substantial interobserver and intraobserver reliability when evaluating the type of diaphyseal femur fractures. Determination of fracture group and subgroup, however, progressively reduces the reliability of the classification, especially for fractures caused by a gunshot. Diaphyseal femur fractures caused by gunshots, by means of their fracture patterns, cannot be classified reliably with the AO/Orthopaedic Trauma Association classification system.

Femoral Fractures↗
PubMed1996

Classification of lung tumors on chest radiographs by fractal texture analysis.

RATIONALE AND OBJECTIVES: The authors determine the usefulness of fractal texture analysis in classification of lung tumors on chest radiographs. METHODS: A method of fractal texture analysis was applied for classification of lung tumors on digitized chest radiographs. It was performed in 15 cases of benign lesions and in 26 cases of malignant tumors. All lesions were proved by histology. For classification, the fractal distances of the tumors were calculated and an energy measure related to the variances as a second characteristic was applied. The results were compared with those of a conventional classification method based on the co-occurrence matrix. RESULTS: Although the conventional method failed in classification of lung tumors, a clear separation between benign and malignant lesions was attained by fractal analysis. A differentiation between primary lung cancer and metastases was not possible. CONCLUSIONS: The results of our study demonstrate the usefulness of fractal texture analysis for classification of lung tumors on chest radiographs.

Fractals↗
PubMed1997

Assessment of the AO/ASIF fracture classification for the distal tibia.

OBJECTIVES: The purpose of this study was to assess the interobserver reliability and intraobserver reproducibility of the AO/ASIF and Rüedi and Allgöwer classifications for fractures of the distal tibia, and to determine the benefit of a computed tomography (CT) scan and experience on observer agreement for several fracture characteristics, including classification. METHODS: The radiographs of forty-three fractures of the distal tibia, fourteen of which had CT scans, were assessed by groups of experienced and less-experienced observers. Each case was classified according to the AO/ASIF and Rüedi and Allgöwer systems. Several other fracture characteristics also were assessed. The kappa coefficient of agreement was calculated and used to compare the interobserver reliability and intraobserver reproducibility of the classification systems and to determine the benefit of experience and CT scans. The intraclass correlation coefficient was used to assess noncategoric data. RESULTS: Interobserver and intraobserver agreements were good when classifying fractures into AO/ASIF types and significantly better than that for the Rüedi and Allgöwer system. However, agreement was poor when classifying the fractures into AO/ASIF groups. For most assessments, the experienced group tended to have higher levels of interobserver agreement, but not intraobserver agreement. Viewing the CT scans improved agreement on the percentage of articular surface involved, but it did not improve interobserver reliability or intraobserver reproducibility for either of the classification systems. CONCLUSION: The AO/ASIF classification for fractures of the distal tibia has good observer agreement at the type level, but poor agreement at the group level. Experience tends to improve interobserver agreement, but not intraobserver agreement. Viewing CT scans does not improve agreement on classification, but it tends to improve agreement on articular surface involvement.

Ankle Injuries↗
PubMed2002

Impact of MRI on treatment plan and fracture classification of tibial plateau fractures.

OBJECTIVE: To evaluate the interobserver agreement for both treatment plan and fracture classification of tibial plateau fractures using plain radiographs, computed tomography (CT) scan, and magnetic resonance imaging (MRI). DESIGN: Prospective study to assess the impact of an advanced radiographic study on the agreement of treatment plan and fracture classification of tibial plateau fractures among three orthopaedic surgeons. SETTING/PARTICIPANTS: Patients presenting with tibial plateau fractures to a level I trauma center were evaluated with plain knee radiographs (anteroposterior, lateral, two oblique views), CT scan, and MRI. Three experienced attending orthopaedic trauma surgeons were randomly presented three sets of studies for each injury: radiographs alone, radiographs with CT, and radiographs with MRI (including soft tissue injuries documented by an experienced MRI radiologist). The surgeons were asked to render fracture classification and treatment plan based upon the blind reading of each individual radiographic set. MAIN OUTCOME MEASURES: Agreement among the three surgeons was measured using kappa coefficients. RESULTS: For fracture classification, radiographs alone yielded a mean kappa coefficient of 0.68, which increased to 0.73 for radiographs with CT scan and 0.85 for radiographs with MRI. Fracture classification (Schatzker) was changed an average of 6% with the addition of the CT scan and 21% based on radiographs with MRI. For the fracture management plan, the mean interobserver kappa coefficient for radiographs alone was 0.72, which increased to 0.77 for radiographs with CT scan and 0.86 for radiographs with MRI. MRI changed treatment plan in 23% of the cases. CONCLUSION: Magnetic resonance imaging increases the interobserver agreement on fracture classification and operative management of tibial plateau fractures.

Adult↗
PubMed2003

Impact of bone density on distal radius fracture patterns and comparison between five different fracture classifications.

OBJECTIVE: To investigate the impact of bone mineral density (BMD) and bone geometry on failure loads and fracture patterns of the distal radius and to compare 5 different fracture classifications. DESIGN: Biomechanical and radiologic in vitro study. SETTING: Research laboratory. MAIN OUTCOME MEASUREMENTS: A total of 118 intact human forearms from elderly donors were examined by means of conventional radiography and peripheral quantitative computed tomography (PQCT) to determine BMD and geometry. The forearms were subjected to a standardized biomechanical test simulating a fall on the outstretched hand. The distal radius fractures were classified from x-rays using the AO ( 33), Cooney ( 9), Fernandez ( 15), Frykman ( 17), and Melone ( 31) classifications. The grading was repeated after preparation and direct visual inspection of the fracture site and was correlated with radiographic results. Fracture patterns also were correlated with BMD and geometry. RESULTS: Correlations between bone properties and fracture patterns (r = 0.09-0.70) suggested an increase in the severity of fractures with decreasing bone quality. The highest correlation between failure load and bone properties was found for the cortical area (r = 0.70) and trabecular density (r = 0.60). Good correlations between radiographic and direct visual classification were obtained for the Cooney ( 9) (r = 0.70), the AO ( 33) (r = 0.68), and the Fernandez ( 15) (r = 0.65) classifications. Smaller values were found for the Frykman ( 17) (r = 0.44) and the Melone ( 31) (r = 0.27) classifications. CONCLUSIONS: With increasing osteopenia, the load to failure decreases, and the severity of fractures increases. Fracture patterns in this patient population can be adequately graded with the AO ( 33) and Cooney ( 9) classifications. The severity of distal radius fractures tends to be underestimated by conventional x-ray examination, which needs to be taken into account when a fracture treatment plan is selected.

Aged↗
PubMed1998

Classification of otitis media.

OBJECTIVES: To review the International Classification of Diseases (ICD) system of classifying otitis media. To highlight the failings of this system and to propose an alternative (Read version 3). STUDY DESIGN: An historical review of the literature and a presentation of two distinct classification schemes. METHODS: The advantages and disadvantages of the alternative schemes are analyzed and an algorithm for the classification of otitis media is presented that describes the decisions taken to reach a diagnosis. RESULTS: The proposed classification of otitis media as used in the Read Codes thesaurus is more logical than the ICD classification. CONCLUSIONS: The widely used ICD schema of otitis media fails to provide the otologist or general medical practitioner with a logically organized set of terms to describe inflammatory middle ear disease. This article proposes a simple, hierarchical classification that can be used by specialists, generalists and epidemiologists alike.

Acute Disease↗
PubMed1982

Three case-type classifications: suitability for use in reimbursing hospitals.

This study compared three case-type classifications--the cross-classification of the Commission on Professional and Hospital Activities, Diagnosis-Related Groups (DRGs) and Staging--with respect to per cent of variance in total patient charges accounted for. The purpose was to assess the relative usefulness of the classifications for application in hospital reimbursement schemes. The sample consisted of 50 hospitals. A nested analysis of variance was performed with case type nested within hospital. Per cent of variance accounted for was calculated for each of three data sets: the full data set, a truncated version of that set and a logarithmically transformed version. Results support the contention that none of the currently available classifications accounts for enough variance to permit straightforward use of case-type standard costs in a reimbursement mechanism. New developments in case-type classification may result in a classification that is more suitable for this use.

Analysis of Variance↗