Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Classification Algorithms”

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 73 records · Page 4Linked to original sources
PubMed1996

Automated topographic screening for keratoconus in refractive surgery candidates.

PURPOSE: We evaluated an automated corneal topography classification system developed as an adjuvant for screening patients prior to keratorefractive surgery. We screened for patterns suspicious for keratoconus by applying the system to the analysis of a series of patients who presented for evaluation for surgical correction of myopia. METHODS: Both eyes of 53 consecutive patients who were included in a previously reported prospective study were evaluated using the Expert System classification algorithm. This quantitative classification system incorporating eight indices was applied to the videokeratoscopic data from each patient to divide the topographic patterns into keratoconus and non-keratoconus groups. The group assignment of the Expert System classifier was compared with the clinical diagnosis of keratoconus versus non-keratoconus based on the topographic pattern and objective biomicroscopy signs. RESULTS: The Expert System classified eight of the videokeratographs as keratoconus. All five corneas that had clinical evidence of keratoconus were classified as such by the Expert System (sensitivity 100%). The other three corneas that were classified as keratoconus were of patients who wore rigid contact lenses and had pseudo-keratoconus topographic patterns, without other clinical signs of keratoconus. The specificity with which the Expert System detected normal corneas was 97% (98/101). CONCLUSIONS: Evaluation of the videokeratographic data with computerized algorithms designed to detect keratoconus may aid preoperative evaluation and facilitate distinction between keratoconus and some keratoconus-like topographic patterns.

Algorithms↗
PubMed2006

Neuroblastoma detection using serum proteomic profiling: a novel mining technique for cancer?

BACKGROUND: Serum proteins in neuroblastoma (NB), such as neuron-specific enolase and lactate dehydrogenase, are used as nonspecific markers of disease severity. In this study, we have generated serum protein profiles that correlate with NB by applying proteomic technologies to uncover, at the protein level, serum polypeptide expression patterns in patients with NB. METHODS: Surface-enhanced laser desorption/ionization-time-of-flight mass spectrometry was used to generate protein expression spectra in human NB (group I, n = 18) and healthy children (group II, n = 17) sera. Groups I and II mass spectral data were compared after baseline subtraction. Peaks with high signal-to-noise ratios were selected and grouped into bins with various intervals along mass-to-charge axis. Two-sample t test and 3-fold cross validation were used to identify differential biomarkers between groups I and II. RESULTS: Significant differentially expressed proteins were identified between groups I and II (P < or = .05). The discriminatory features (proteomic patterns) of cancer from normal sera were successfully identified using the classification algorithm. The average classification performance after 3-fold cross validation was 87.26%. CONCLUSION: Surface-enhanced laser desorption/ionization-time-of-flight mass spectrometry is suitable for preliminary assessment of NB and could potentially provide a noninvasive diagnosis of NB. We propose that surface-enhanced laser desorption/ionization provides a novel method for NB diagnosis because direct observations of spectral differences between normal and NB sera are possible.

Child, Preschool↗
PubMedDate not supplied

Prevalence of irritable bowel syndrome (IBS) and variability of diagnostic criteria.

OBJECTIVES: The main objectives of this study were to assess whether the use of different definitions of irritable bowel syndrome (IBS) could influence measurements of its prevalence and characterize the patient population fulfilling these different diagnostic criteria. METHOD: A telephone survey was carried out by contacting 8,221 subjects aged >or=18 Years representative of the French population. A "screening" questionnaire based on three algorithms of IBS classification (Manning, with or without a notion of a minimal duration of symptoms, Rome I and Rome II) was used by specialised inquirers. RESULTS: Twenty three percent of the subjects interviewed stated that they had suffered from abdominal pain during the previous 12 Months. The prevalence of IBS considerably varied, depending on the diagnostic criteria used: 12% based on Manning criteria without reference to the duration of symptoms; 2.5% if the notion of duration of symptoms was added to the Manning criteria, and 2.1% and 1.1% based on the Rome I and Rome II criteria, respectively (the latter including the same notion of duration). In total, 212 subjects (2.6%) met at least one of the criteria including a minimal duration of symptoms, with a predominance for women (sex-ratio close to 2). CONCLUSION: The prevalence of IBS is strongly dependent on the classification algorithm employed. The requirement of a minimum duration of symptoms eliminates IBS in a large number of subjects complaining of abdominal disorders. Once these methodological variations were taken into account, the prevalence of IBS in France was found to be comparable to that published in international literature.

Abdominal Pain↗
PubMed2006

Searching for biomarkers of heart failure in the mass spectra of blood plasma.

We have developed a technique for analysing blood plasma using MALDI-MS with subsequent data analysis to identify significant and specific differences between heart failure (HF) patients and healthy individuals. A training dataset comprising 100 HF patients and 100 healthy individuals was used to search for biomarkers (m/z range 1000-10,000). EWP cartridges when used in tandem with microcon centrifugal filters were found to give the best results. A data management chain including event binning, background subtraction and feature extraction was developed to reduce the data, and statistical analysis was used to map feature intensities on to a common scale. Various mathematical approaches including a simple cumulative score, support vector machines (SVM) and genetic algorithms (GAs) were then used to combine the results from individual features and provide a robust classification algorithm. The SVM gave the most promising results (accuracy 95%, receiver operating characteristic (ROC) score of 0.997 using 18 selected features). Finally, a test dataset comprising a further 32 HF patients and 20 controls was used to verify that the 18 putative biomarkers and classification algorithms gave reliable predictions (accuracy 88.5%, ROC score 0.998).

Biomarkers↗
PubMed2004

An efficient rhythmic component expression and weighting synthesis strategy for classifying motor imagery EEG in a brain-computer interface.

The recognition of mental states during motor imagery tasks is crucial for EEG-based brain-computer interface research. We have developed a new algorithm by means of frequency decomposition and weighting synthesis strategy for recognizing imagined right- and left-hand movements. A frequency range from 5 to 25 Hz was divided into 20 band bins for each trial, and the corresponding envelopes of filtered EEG signals for each trial were extracted as a measure of instantaneous power at each frequency band. The dimensionality of the feature space was reduced from 200 (corresponding to 2 s) to 3 by down-sampling of envelopes of the feature signals, and subsequently applying principal component analysis. The linear discriminate analysis algorithm was then used to classify the features, due to its generalization capability. Each frequency band bin was weighted by a function determined according to the classification accuracy during the training process. The present classification algorithm was applied to a dataset of nine human subjects, and achieved a success rate of classification of 90% in training and 77% in testing. The present promising results suggest that the present classification algorithm can be used in initiating a general-purpose mental state recognition based on motor imagery tasks.

Algorithms↗
PubMed1981

Classification of gynecologic flow cytometry data: a comparison of methods.

Several discriminant function methods for automatically classifying flow cytometry data from human cervical material were developed and compared with previously published methods using a sample of 186 specimens. The misclassification rates (approximately 20%) were similar to those of other published techniques for classifying these data. The methods misclassify different cases, however. The apparent system performance appears to be limited by at least three factors: (1) use of too small a sample in constructing classification algorithms, (2) poor "visibility" of small numbers of abnormal cells in the flow histograms and (3) incorrect or inconsistent visual classification of the samples used to construct the classification algorithms. The third factor results in erroneously high estimates of the misclassification rate. Even so, the overall system performance appears to be comparable to that of many cytotechnologists.

Cervix Uteri↗
PubMed2005

An incremental approach to genetic-algorithms-based classification.

Incremental learning has been widely addressed in the machine learning literature to cope with learning tasks where the learning environment is ever changing or training samples become available over time. However, most research work explores incremental learning with statistical algorithms or neural networks, rather than evolutionary algorithms. The work in this paper employs genetic algorithms (GAs) as basic learning algorithms for incremental learning within one or more classifier agents in a multiagent environment. Four new approaches with different initialization schemes are proposed. They keep the old solutions and use an "integration" operation to integrate them with new elements to accommodate new attributes, while biased mutation and crossover operations are adopted to further evolve a reinforced solution. The simulation results on benchmark classification data sets show that the proposed approaches can deal with the arrival of new input attributes and integrate them with the original input space. It is also shown that the proposed approaches can be successfully used for incremental learning and improve classification rates as compared to the retraining GA. Possible applications for continuous incremental training and feature selection are also discussed.

Algorithms↗
PubMed2003

A novel approach to describe a U1 snRNA binding site.

RNA duplex formation between U1 snRNA and a splice donor (SD) site can protect pre-mRNA from degradation prior to splicing and initiates formation of the spliceosome. This process was monitored, using sub-genomic HIV-1 expression vectors, by expression analysis of the glycoprotein env, whose formation critically depends on functional SD4. We systematically derived a hydrogen bond model for the complementarity between the free 5' end of U1 snRNA and 5' splice sites and numerous mutations following transient transfection of HeLa-T4+ cells with 5' splice site mutated vectors. The resulting model takes into account number, interdependence and neighborhood relationships of predicted hydrogen bond formation in a region spanning the three most 3' base pairs of the exon (-3 to -1) and the eight most 5' base pairs of the intron (+1 to +8). The model is represented by an algorithm classifying U1 snRNA binding sites which can or cannot functionally substitute SD4 with respect to Rev-mediated env expression. In a data set of 5' splice site mutations of the human ATM gene we found a significant correlation between the algorithmic classification and exon skipping (P = 0.018, chi2-test), showing that the applicability of the proposed model reaches far beyond HIV-1 splicing. However, the algorithmic classification must not be taken as an absolute measure of SD usage as it may be modified by upstream sequence elements. Upstream to SD4 we identified a fragment supporting ASF/SF2 binding. Mutating GAR nucleotide repeats within this site decreased the SD4-dependent Rev-mediated env expression, which could be balanced simply by artificially increasing the complementarity of SD4.

Algorithms↗
PubMed2005

A simple prediction algorithm for bacteraemia in patients with acute febrile illness.

BACKGROUND: Existing prediction models for the risk of bacteraemia are complex and difficult to use. Physicians are likely to use a model only if it is simple and sensitive. AIM: To develop a simple classification algorithm predicting the risk of bacteraemia. DESIGN: Hospital-based study. METHODS: We enrolled 526 adult consecutive patients with acute febrile illness (40 with bacteraemia) presenting to the emergency department at a community hospital in Okinawa, Japan. Recursive partitioning analysis was used to build the classification algorithm with V-fold cross-validation. We used two clinical scenarios: in the first, laboratory tests were not available; in the second, they were. RESULTS: The two prediction algorithms generated three different risk groups for bacteraemia. In the first scenario, the important variables were chills, pulse, and physician diagnosis of a low-risk site. The low-risk group from this first algorithm included 68% of the total patients; sensitivity was 87.5% and the misclassification rate was 1.4% (5/358). In the second scenario, the important variables were chills, C-reactive protein, and physician diagnosis of a low-risk site. The low-risk group for the second algorithm included 62% of the total patients; sensitivity was 92.5% and misclassification rate 0.9% (3/328). The algorithms had negative predictive values of 98.6% (first scenario) and 99.1% (second). DISCUSSION: This simple and sensitive prediction algorithm may be useful for identifying patients at low risk of bacteraemia. Prospective validation is needed in other settings.

Acute Disease↗
PubMed2004

Investigation of fiber-optic probe designs for optical spectroscopic diagnosis of epithelial pre-cancers.

BACKGROUND AND OBJECTIVES: The first objective of this study was to evaluate the performance of fluorescence spectroscopy for diagnosing pre-cancers in stratified squamous epithelial tissues in vivo using two different probe geometries with (1) overlapping versus (2) non-overlapping illumination and collection areas on the tissue surface. Probe (1) and probe (2) are preferentially sensitive to the fluorescence originating from the tissue surface and sub-surface tissue depths, respectively. The second objective was to design a novel, angled illumination fiber-optic probe to maximally exploit the depth-dependent fluorescence properties of epithelial tissues. STUDY DESIGN/MATERIALS AND METHODS: In the first study, spectra were measured from epithelial pre-cancers and normal tissues in the hamster cheek pouch and analyzed with a non-parametric classification algorithm. In the second study, Monte Carlo modeling was used to simulate fluorescence measurements from an epithelial tissue model with the angled illumination probe. RESULTS: An unbiased classification algorithm based on spectra measured with probes (1) and (2), classified pre-cancerous and normal tissues with 78 and 94% accuracy, respectively. The angled illumination probe design provides the capability to detect fluorescence from a wide range of tissue depths in an epithelial tissue model. CONCLUSIONS: The first study demonstrates that fluorescence originating from sub-surface tissue depths (probe (2)) is more diagnostic than fluorescence originating from the tissue surface (probe (1)) in the hamster cheek pouch model. However in general, it is difficult to know a priori the optimal probe geometry for pre-cancer detection in a particular epithelial tissue model. The angled illumination probe provides the capability to measure tissue fluorescence selectively from different depths within epithelial tissues, thus obviating the need to select a single optimal probe design for the fluorescence-based diagnosis of epithelial pre-cancers.

Algorithms↗
PubMed1976

Automatic leukocyte classification using cytochemically stained smears.

A leukocyte classification algorithm suitable for automated differential counting has been developed for blood smears stained with a new three-component cytochemical stain which has relatively narrow absorption bands centered at 460, 540 and 640 nm, respectively. The classification procedure is the result of a pattern recognition experiment using a sample of 223 leukocytes distributed evenly over the five normal cell types. The basic data for each cell were three digital microscopic images obtained with narrow band illumination at the above central wavelengths using a TV-digitizer system interfaced to a PDP-15 computer. The classification algorithm involves a sequential decision procedure utilizing five pattern features computed from the intensity histograms of the green and blue digital images. Thus the number of arithmetic operations and the number of computer memory words necessary to perform the classification into one of the five normal white blood cell types are both proportional to n where n is the number of gray levels into which the intensity scale is divided. In this experiment, n equals 256. Comparison of our results with work of others on smears prepared with Romanowski-type stains indicates that such narrow-band, spectrally well separated cytochemical multiple stains can permit the use of algorithms which are approximately ten times faster.

Autoanalysis↗
PubMed2005

A genotype calling algorithm for affymetrix SNP arrays.

MOTIVATION: A classification algorithm, based on a multi-chip, multi-SNP approach is proposed for Affymetrix SNP arrays. Current procedures for calling genotypes on SNP arrays process all the features associated with one chip and one SNP at a time. Using a large training sample where the genotype labels are known, we develop a supervised learning algorithm to obtain more accurate classification results on new data. The method we propose, RLMM, is based on a robustly fitted, linear model and uses the Mahalanobis distance for classification. The chip-to-chip non-biological variance is reduced through normalization. This model-based algorithm captures the similarities across genotype groups and probes, as well as across thousands of SNPs for accurate classification. In this paper, we apply RLMM to Affymetrix 100 K SNP array data, present classification results and compare them with genotype calls obtained from the Affymetrix procedure DM, as well as to the publicly available genotype calls from the HapMap project.

Algorithms↗
PubMed2004

Automated detection of obstructive sleep apnoea at different time scales using the electrocardiogram.

An automated classification algorithm is presented which processes short-duration epochs of surface electrocardiogram data derived from polysomnography studies, and determines whether an epoch is from a period of sleep disordered respiration (SDR) or normal respiration (NR). The epoch lengths considered were 15, 30, 45, 60, 75, and 90 s. Epochs were labeled as 'NR' or 'SDR' by a human expert, based on standard polysomnography interpretation rules. The automated classification algorithm was trained and tested on a database of 70 overnight ECG recordings from subjects with and without obstructive sleep apnoea syndrome (35 used for training, 35 for independent validation). Depending on the epoch length, the classifier correctly labeled between 87% (15 s epochs) and 91% (60 s epochs) of the epochs in the test set. Accuracy was lowest for the shortest (15 s) and longest (90 s) epoch lengths, but the analysis was relatively insensitive to choice of epoch length. The classifications from these epochs were combined to form an overall summary measure of minutes-of-SDR, allowing per-subject classification.

Adult↗
PubMed2003

Antemortem identification of bovine spongiform encephalopathy from serum using infrared spectroscopy.

Since 1986, more than 180 000 clinical cases of bovine spongiform encephalopathy (BSE) have been observed in the U.K. alone. Most of these cases were confirmed by postmortem examination of brain tissue. However, BSE-related risk assessment and risk management would greatly benefit from antemortem testing on living animals. A serum-based test could allow for screening of the cattle population; thus, even a BSE eradication program would be conceivable. Here we report on a novel method for antemortem BSE testing, which combines infrared spectroscopy of serum samples with multivariate pattern recognition analysis. A classification algorithm was trained using infrared spectra of bovine sera from more than 800 animals (including BSE-positive, healthy controls and animals suffering from classical viral or bacterial infections). In two validation studies, sensitivities of 85 and 84% and specificities of 86 and 91% were achieved, respectively. The combination of classification algorithms increased the sensitivity and specificity of BSE detection to 96 and 92%, respectively.

Animals↗
PubMed2006

Application of surface-enhanced laser desorption/ionization time-of-flight-based serum proteomic array technique for the early diagnosis of prostate cancer.

AIM: To identify the serum biomarkers of prostate cancer (PCa) by protein chip and bioinformatics. METHODS: Serum samples from 83 PCa patients and 95 healthy men were taken from a mass screening in Changchun, China. Protein profiling was carried out using surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF MS). The data of spectra were analyzed using two bioinformatics tools. RESULTS: Eighteen serum differential proteins were identified in the PCa group compared with the control group (P < 0.01). There were four proteins at the higher serum level and 14 proteins at the lower serum level in the PCa group. A decision tree classification algorithm that used an eight-protein mass pattern was developed to correctly classify the samples. A sensitivity of 92.0% and a specificity of 96.7% for the study group were obtained by comparing the PCa and control groups. CONCLUSION: We identified new serum biomarkers of PCa. SELDI-TOF MS coupled with a decision tree classification algorithm will provide a highly accurate and innovative approach for the early diagnosis of PCa.

Aged↗
PubMed2005

Feature selection and nearest centroid classification for protein mass spectrometry.

BACKGROUND: The use of mass spectrometry as a proteomics tool is poised to revolutionize early disease diagnosis and biomarker identification. Unfortunately, before standard supervised classification algorithms can be employed, the "curse of dimensionality" needs to be solved. Due to the sheer amount of information contained within the mass spectra, most standard machine learning techniques cannot be directly applied. Instead, feature selection techniques are used to first reduce the dimensionality of the input space and thus enable the subsequent use of classification algorithms. This paper examines feature selection techniques for proteomic mass spectrometry. RESULTS: This study examines the performance of the nearest centroid classifier coupled with the following feature selection algorithms. Student-t test, Kolmogorov-Smirnov test, and the P-test are univariate statistics used for filter-based feature ranking. From the wrapper approaches we tested sequential forward selection and a modified version of sequential backward selection. Embedded approaches included shrunken nearest centroid and a novel version of boosting based feature selection we developed. In addition, we tested several dimensionality reduction approaches, namely principal component analysis and principal component analysis coupled with linear discriminant analysis. To fairly assess each algorithm, evaluation was done using stratified cross validation with an internal leave-one-out cross-validation loop for automated feature selection. Comprehensive experiments, conducted on five popular cancer data sets, revealed that the less advocated sequential forward selection and boosted feature selection algorithms produce the most consistent results across all data sets. In contrast, the state-of-the-art performance reported on isolated data sets for several of the studied algorithms, does not hold across all data sets. CONCLUSION: This study tested a number of popular feature selection methods using the nearest centroid classifier and found that several reportedly state-of-the-art algorithms in fact perform rather poorly when tested via stratified cross-validation. The revealed inconsistencies provide clear evidence that algorithm evaluation should be performed on several data sets using a consistent (i.e., non-randomized, stratified) cross-validation procedure in order for the conclusions to be statistically sound.

Algorithms↗
PubMed2006

Protein solubility: sequence based prediction and experimental verification.

MOTIVATION: Obtaining soluble proteins in sufficient concentrations is a recurring limiting factor in various experimental studies. Solubility is an individual trait of proteins which, under a given set of experimental conditions, is determined by their amino acid sequence. Accurate theoretical prediction of solubility from sequence is instrumental for setting priorities on targets in large-scale proteomics projects. RESULTS: We present a machine-learning approach called PROSO to assess the chance of a protein to be soluble upon heterologous expression in Escherichia coli based on its amino acid composition. The classification algorithm is organized as a two-layered structure in which the output of primary support vector machine (SVM) classifiers serves as input for a secondary Naive Bayes classifier. Experimental progress information from the TargetDB database as well as previously published datasets were used as the source of training data. In comparison with previously published methods our classification algorithm possesses improved discriminatory capacity characterized by the Matthews Correlation Coefficient (MCC) of 0.434 between predicted and known solubility states and the overall prediction accuracy of 72% (75 and 68% for positive and negative class, respectively). We also provide experimental verification of our predictions using solubility measurements for 31 mutational variants of two different proteins.

Amino Acid Sequence↗
PubMedDate not supplied

Laguerre-based method for analysis of time-resolved fluorescence data: application to in-vivo characterization and diagnosis of atherosclerotic lesions.

We report the application of the Laguerre deconvolution technique (LDT) to the analysis of in-vivo time-resolved laser-induced fluorescence spectroscopy (TR-LIFS) data and the diagnosis of atherosclerotic plaques. TR-LIFS measurements were obtained in vivo from normal and atherosclerotic aortas (eight rabbits, 73 areas), and subsequently analyzed using LDT. Spectral and time-resolved features were used to develop four classification algorithms: linear discriminant analysis (LDA), stepwise LDA (SLDA), principal component analysis (PCA), and artificial neural network (ANN). Accurate deconvolution of TR-LIFS in-vivo measurements from normal and atherosclerotic arteries was provided by LDT. The derived Laguerre expansion coefficients reflected changes in the arterial biochemical composition, and provided a means to discriminate lesions rich in macrophages with high sensitivity (>85%) and specificity (>95%). Classification algorithms (SLDA and PCA) using a selected number of features with maximum discriminating power provided the best performance. This study demonstrates the potential of the LDT for in-vivo tissue diagnosis, and specifically for the detection of macrophages infiltration in atherosclerotic lesions, a key marker of plaque vulnerability.

Algorithms↗