Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Feature selection”

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 829 records · Page 46Linked to original sources
PubMed2005

A machine learning approach for automated recognition of movement patterns using basic, kinetic and kinematic gait data.

This paper investigated application of a machine learning approach (Support vector machine, SVM) for the automatic recognition of gait changes due to ageing using three types of gait measures: basic temporal/spatial, kinetic and kinematic. The gaits of 12 young and 12 elderly participants were recorded and analysed using a synchronized PEAK motion analysis system and a force platform during normal walking. Altogether, 24 gait features describing the three types of gait characteristics were extracted for developing gait recognition models and later testing of generalization performance. Test results indicated an overall accuracy of 91.7% by the SVM in its capacity to distinguish the two gait patterns. The classification ability of the SVM was found to be unaffected across six kernel functions (linear, polynomial, radial basis, exponential radial basis, multi-layer perceptron and spline). Gait recognition rate improved when features were selected from different gait data type. A feature selection algorithm demonstrated that as little as three gait features, one selected from each data type, could effectively distinguish the age groups with 100% accuracy. These results demonstrate considerable potential in applying SVMs in gait classification for many applications.

Aged↗
PubMed2000

Neuronal mechanisms of selectivity for object features revealed by blocking inhibition in inferotemporal cortex.

The inferotemporal cortex (area TE) of monkeys, a higher station of the visual information stream for object recognition, contains neurons selective for particular object features. Little is known about how and where this selectivity is generated. We show that blockade of inhibition mediated by gamma-aminobutyric acid (GABA) markedly altered the selectivity of TE neurons by augmenting their responses to some stimuli but not to others. The effects were observed for particular groups of stimuli related to the originally effective stimuli or those that did not originally excite the neurons but activated nearby neurons. Intrinsic neuronal interactions within area TE thus determine the final characteristic of their selectivity, and GABAergic inhibition contributes to this process.

Animals↗
PubMed2026

Robust prioritization of genomic features with stability selection.

MOTIVATION: The heterogeneity of complex diseases including cancer leads to heavy-tailed distributions in the disease traits. In such settings, non-robust variable selection methods are inherently susceptible to data contamination and can yield unstable or misleading results. This vulnerability becomes more severe for recently proposed approaches that introduce pseudo-features as negative controls, as these methods further amplify the curse of dimensionality by expanding the genotype matrix in the presence of outliers and high-dimensional genomic features. RESULTS: We develop a robust variable selection framework with stability selection to prioritize genomic features in the presence of contamination. In contrast to existing approaches that rely on pseudo-features for error control, the proposed method achieves double robustness. First, it adopts least absolute deviation (LAD) LASSO to ensure robustness against outliers and heavy-tailed errors in disease traits. Second, it avoids augmenting the genotype matrix with pseudo-features, thereby mitigating the curse of dimensionality that is particularly problematic in high-dimensional genomic data. The proposed method has been extensively evaluated in simulation studies to demonstrate its effectiveness over multiple competing methods for variable selection. In addition, we have applied the proposed method and competing approaches to two real-data case studies: the The Cancer Genome Atlas (TCGA) Skin Cutaneous Melanoma (SKCM) dataset and an eQTL dataset. The results demonstrate that the proposed method achieves superior performance by identifying genomic features with higher reproducibility. AVAILABILITY AND IMPLEMENTATION: The source code for implementing the proposed methods is publicly available at https://github.com/cenwu/RSS with an archival DOI https://doi.org/10.6084/m9.figshare.32306883.

Genomics↗
PubMed1999

Attentional selection of objects or features: evidence from a modified search task.

Three experiments examined the domain of visual selective attention (i.e., feature-based selection vs. object-based selection). Experiment 1 extended the requirements of the visual search task by requiring a feature discrimination response to target elements presented for short durations (30-105 msec). Targets were embedded in 47 distractor elements and were defined by either a distinct color or a distinct orientation. Observers made a discrimination response to either the target's color or its orientation. When the target-defining feature and the feature to be discriminated were the same (matched conditions), accuracy was enhanced relative to when these features belonged to separate dimensions (mismatched conditions). In Experiment 2, similar results were found in a task in which the target-defining dimension varied from trial to trial and observers performed both color and orientation discriminations on every trial. The results from these two experiments are consistent with feature-based attentional selection, but not with object-based selection. Experiment 3 extended these findings by showing that the effect is rooted in the overlap between target and distractor values in the stimulus set. The results are discussed in the context of recent models of visual selective attention.

Attention↗
PubMed1999

Perceiving spatially inseparable objects: evidence for feature-based object selection not mediated by location.

In 4 experiments, stimulus elements were arranged into an LED-like array, and letters were defined within the array by feature similarity between the elements with respect to color and form. These stimuli allowed the display of a target and a distractor letter simultaneously at the same location. They were spatially inseparable but could be separated in feature space. Participants had to identify the letter on a prespecified feature dimension (color or form). As a result, the distractors produced specific compatibility effects. This showed that nontarget features could not be ignored at an early stage (i.e., that color and form were processed automatically and in parallel up to a high stage). The target was selected from the resulting objects according to the prespecified feature dimension. Results demonstrate that object selection is possible without selecting absolute spatial arrays.

Adult↗
PubMed1998

Potent and selective thrombin inhibitors featuring hydrophobic, basic P3-P4-aminoalkyllactam moieties.

Crystal structure and evolving SAR considerations of potent, selective benzylsulfonamide lactam thrombin inhibitors and related serine protease inhibitors have led to the design of novel thrombin inhibitors 1a-g, featuring hydrophobic, basic, P4-alkylaminolactam scaffolds that serve as novel types of P3-P4 dipeptide mimics. The design, synthesis, and biological activity of these targets is presented.

Antithrombins↗
PubMed1989

Aging, selective attention, and feature integration.

This study used feature-integration theory as a means of determining the point in processing at which selective attention deficits originate. The theory posits an initial stage of processing in which features are registered in parallel and then a serial process in which features are conjoined to form complex stimuli. Performance of young and older adults on feature versus conjunction search is compared. Analyses of reaction times and error rates suggest that elderly adults in addition to young adults, can capitalize on the early parallel processing stage of visual information processing, and that age decrements in visual search arise as a result of the later, serial stage of processing. Analyses of a third, unconfounded, conjunction search condition reveal qualitatively similar modes of conjunction search in young and older adults. The contribution of age-related data limitations is found to be secondary to the contribution of age decrements in selective attention.

Adult↗
PubMed1998

3-Hydroxykynurenine, an endogenous oxidative stress generator, causes neuronal cell death with apoptotic features and region selectivity.

3-Hydroxykynurenine (3-HK) is a potential endogenous neurotoxin whose increased levels have been described in several neurodegenerative disorders. Here, we characterized in vitro neurotoxicity of 3-HK. Of the tested kynurenine pathway metabolites, only 3-HK, and to a lesser extent 3-hydroxyanthranilic acid, were toxic to primary cultured striatal neurons. 3-HK toxicity was inhibited by various antioxidants, indicating that the generation of reactive oxygen species is essential to the toxicity. 3-HK-induced neuronal cell death showed several features of apoptosis, as determined by the blockade by macromolecule synthesis inhibitors, and by the observation of cell body shrinkage with nuclear chromatin condensation and fragmentation. In addition, 3-HK toxicity was dependent on its cellular uptake via transporters for large neutral amino acids, because uptake inhibition blocked the toxicity. Cortical and striatal neurons were much more vulnerable to 3-HK toxicity than cerebellar neurons, which may be attributable to the differences in transporter activities of these neurons. These results indicate that 3-HK, depending on transporter-mediated cellular uptake and on intracellular generation of oxidative stress, induces neuronal cell death with brain region selectivity and with apoptotic features, which may be relevant to pathology of neurodegenerative disorders.

Amino Acid Transport Systems↗
PubMed2003

SamCluster: an integrated scheme for automatic discovery of sample classes using gene expression profile.

MOTIVATION: Feature (gene) selection can dramatically improve the accuracy of gene expression profile based sample class prediction. Many statistical methods for feature (gene) selection such as stepwise optimization and Monte Carlo simulation have been developed for tissue sample classification. In contrast to class prediction, few statistical and computational methods for feature selection have been applied to clustering algorithms for pattern discovery. RESULTS: An integrated scheme and corresponding program SamCluster for automatic discovery of sample classes based on gene expression profile is presented in this report. The scheme incorporates the feature selection algorithms based on the calculation of CV (coefficient of variation) and t-test into hierarchical clustering and proceeds as follows. At first, the genes with their CV greater than the pre-specified threshold are selected for cluster analysis, which results in two putative sample classes. Then, significantly differentially expressed genes in the two putative sample classes with p-values < or = 0.01, 0.05, or 0.1 from t-test are selected for further cluster analysis. The above processes were iterated until the two stable sample classes were found. Finally, the consensus sample classes are constructed from the putative classes that are derived from the different CV thresholds, and the best putative sample classes that have the minimum distance between the consensus classes and the putative classes are identified. To evaluate the performance of the feature selection for cluster analysis, the proposed scheme was applied to four expression datasets COLON, LEUKEMIA72, LEUKEMIA38, and OVARIAN. The results show that there are only 5, 1, 0, and 0 samples that have been misclassified, respectively. We conclude that the proposed scheme, SamCluster, is an efficient method for discovery of sample classes using gene expression profile. AVAILABILITY: The related program SamCluster is available upon request or from the web page http://www.sph.uth.tmc.edu:8052/hgc/Downloads.asp.

Algorithms↗
PubMed1998

Reproducing the natural evolution of protein structural features with the selectively infective phage (SIP) technology. The kink in the first strand of antibody kappa domains.

The beta-sandwich structure of immunoglobulin variable domains is characterized by a typical kink in the first strand, which allows the first part of the strand to hydrogen bond to the outer beta-sheet (away from the VH-VL interface) and the second part to the inner beta-sheet. This kink differs in length and sequence between the Vkappa, Vlambda and VH domains and yet is involved in several almost perfectly conserved interactions with framework residues. We have used the selectively infective phage (SIP) system to select the optimal kink region from several defined libraries, using an anti-hemagglutinin single-chain Fv (scFv) fragment as a model system. Both for the kink with the Vkappa domain length and that with the Vlambda length, a sequence distribution was selected that coincides remarkably well with the sequence distribution of natural antibodies. The selected scFv fragments were purified and characterized, and thermodynamic stability was found to be the prime factor responsible for selection. These data show that the SIP technology can be used for optimizing protein structural features by evolutionary approaches.

Bacteriophages↗
PubMedDate not supplied

The selection of grammatical features in word production: the case of plural nouns in German.

Two experiments investigate the effect of number congruency using picture-word interference. Native German participants were required to name pictures of single objects (Nase 'nose') or two instances of the same object (Nasen 'noses') while ignoring simultaneously presented distractor words. Distractor words either had the same number or were different in number. In addition, the type of plural formation (same or different inflectional plural suffix) and the semantic relationship (same or different semantic category) between target and distractor were varied in Experiments 1 and 2. Results showed no effect of number congruency in either experiment. Furthermore, the type of inflectional suffix did not exert an influence on naming latencies in Experiment 1, but semantic relationship led to a significant interference effect in Experiment 2. The results indicate that selection of the number feature diacritic in noun production is not a competitive process. The implications of the results for models of lexical access are discussed.

Humans↗
PubMed2005

Gene selection from microarray data for cancer classification--a machine learning approach.

A DNA microarray can track the expression levels of thousands of genes simultaneously. Previous research has demonstrated that this technology can be useful in the classification of cancers. Cancer microarray data normally contains a small number of samples which have a large number of gene expression levels as features. To select relevant genes involved in different types of cancer remains a challenge. In order to extract useful gene information from cancer microarray data and reduce dimensionality, feature selection algorithms were systematically investigated in this study. Using a correlation-based feature selector combined with machine learning algorithms such as decision trees, naïve Bayes and support vector machines, we show that classification performance at least as good as published results can be obtained on acute leukemia and diffuse large B-cell lymphoma microarray data sets. We also demonstrate that a combined use of different classification and feature selection approaches makes it possible to select relevant genes with high confidence. This is also the first paper which discusses both computational and biological evidence for the involvement of zyxin in leukaemogenesis.

Algorithms↗
PubMed2004

Gene-expression profile changes correlated with tumor progression and lymph node metastasis in esophageal cancer.

PURPOSE: The purpose of this research was to identify molecular clues to tumor progression and lymph node metastasis in esophageal cancer and to test their value as predictive markers. EXPERIMENTAL DESIGN: We explored the gene expression profiles in cDNA array data of a 36-tissue training set of esophageal squamous cell carcinoma (ESCC) by using generalized linear model-based regression analysis and a feature subset selection algorithm. By applying the identified optimal feature sets (predictive gene sets), we trained and developed ensemble classifiers consisting of multiple probabilistic neural networks combined with AdaBoosting to predict tumor stages and lymph node metastasis. We validated the classifier abilities with 18 independent cases of ESCC. RESULTS: We identified 71 genes of 1289 cancer-related genes of which the expression correlated with tumor stages. Of the 71 genes, 47 significantly differed between the Tumor-Node-Metastasis pT1/2 and pT3/4 stages. Cell cycle regulators and transcriptional factors possibly promoting the growth of tumor cells were highly expressed in the early stages of ESCC, whereas adhesion molecules and extracellular matrix-related molecules possibly promoting invasiveness increased in the later stages. For lymph node metastasis, we identified 44 genes with predictive values, which included cell adhesion molecules and cell membrane receptors showing higher expression in node-positive cases and cell cycle regulators and intracellular signaling molecules showing higher expression in node-negative cases. The ensemble classifiers trained with the selected features predicted tumor stage and lymph node metastasis in the 18 validation cases with respective accuracies of 94.4% and 88.9%. This demonstrated the reproducibility and predictive value of the identified features. CONCLUSION: We suggest that these characteristic genes will provide useful information for understanding the malignant nature of ESCC as well as information useful for personalizing the treatments.

Algorithms↗
PubMed1981

Caprine arthritis-encephalitis: clinical features and presence of antibody in selected goat populations.

Features of caprine arthritis-encephalitis, a retrovirus disease of domestic goats, were studied in 60 goats over a 10-year period. The rate of progression and the severity of the disease process were highly variable within and among animals, but the most salient features were chronically swollen joints and bursae, lameness, weight loss, poor coat, mineralization of soft tissue, and death. Of 1,160 goat serum samples from 24 states tested by the immunodiffusion technique, 81% were positive for antibody to caprine arthritis-encephalitis virus antigens.

Animals↗
PubMed2005

Dental photography systems: required features for equipment selection.

Photographic images are becoming a standard of care for visual communication and medicolegal documentation in contemporary dentistry. For visual images to be effective in diagnosis and treatment planning they must capture useful visual information. Repeatable and predictable image alignment, magnification, and exposure are required for predictable macrophotography in dental practice. This article discusses basic photographic principles that influence specific dental imaging needs and the camera system features required to achieve repeatable diagnostic results.

Humans↗
PubMed2000

Aging and the development of automaticity in conjunction search.

In two experiments, younger and older observers carried out feature searches for targets defined by their luminance contrast and orientation. Additionally, they received consistent-mapping (CM) training in luminance contrast by orientation conjunction search, followed by a brief exposure to conjunction search under reversal conditions. In Experiment 1, display size effects on reaction time suggested that both younger and older observers were conducting a parallel search in all conditions and showed equivalent disruption at reversal. Experiment 2 was a substantive replication of the first using more difficult conjunction search displays. In addition to latency, we measured the number, duration, and feature-based selectivity of fixations made during conjunction search. Display size effects were larger than in Experiment 2 and were of equivalent magnitude in younger and older people. There were no age differences in improvement in conjunction search and minimal age differences in disruption following reversal. Both age groups demonstrated early in training that they could select items possessing target features (i.e., the color white), and both age groups demonstrated that they could not completely reverse this selectivity when these features no longer defined the target. These experiments have several implications for models of visual attention and age differences therein.

Adult↗
PubMedDate not supplied

Saccade target selection in macaque during feature and conjunction visual search.

To gain insight into how vision guides eye movements, monkeys were trained to make a single saccade to a specified target stimulus during feature and conjunction search with stimuli discriminated by color and shape. Monkeys performed both tasks at levels well above chance. The latencies of saccades to the target in conjunction search exhibited shallow positive slopes as a function of set size, comparable to slopes of reaction time of humans during target present/absent judgments, but significantly different than the slopes in feature search. Properties of the selection process were revealed by the occasional saccades to distractors. During feature search, errant saccades were directed more often to a distractor near the target than to a distractor at any other location. In contrast, during conjunction search, saccades to distractors were guided more by similarity than proximity to the target; monkeys were significantly more likely to shift gaze to a distractor that had one of the target features than to a distractor that had none. Overall, color and shape information were used to similar degrees in the search for the conjunction target. However, in single sessions we observed an increased tendency of saccades to a distractor that had been the target in the previous experimental session. The establishment of this tendency across sessions at least a day apart and its persistence throughout a session distinguish this phenomenon from the short-term (<10 trials) perceptual priming observed in this and earlier studies using feature visual search. Our findings support the hypothesis that the target in at least some conjunction visual searches can be detected efficiently based on visual similarity, most likely through parallel processing of the individual features that define the stimuli. These observations guide the interpretation of neurophysiological data and constrain the development of computational models.

Animals↗
PubMed2003

Ultrasonic liver tissues classification by fractal feature vector based on M-band wavelet transform.

This paper describes the feasibility of selecting fractal feature vector based on M-band wavelet transform to classify ultrasonic liver images-normal liver, cirrhosis, and hepatoma. The proposed feature extraction algorithm is based on the spatial-frequency decomposition and fractal geometry. Various classification algorithms based on respective texture measurements and filter banks are presented and tested. Classifications for the three sets of ultrasonic liver images reveal that the fractal feature vector based on M-band wavelet transform is trustworthy. A hierarchical classifier, which is based on the proposed feature extraction algorithm is at least 96.7% accurate in the distinction between normal and abnormal liver images and is at least 93.6% accurate in the distinction between cirrhosis and hepatoma liver images. Additionally, the criterion for feature selection is specified and employed for performance comparisons herein.

Algorithms↗