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 487 records · Page 27Linked to original sources
PubMedDate not supplied

Assessing different classification methods for virtual screening.

How well do different classification methods perform in selecting the ligands of a protein target out of large compound collections not used to train the model? Support vector machines, random forest, artificial neural networks, k-nearest-neighbor classification with genetic-algorithm-optimized feature selection, trend vectors, naïve Bayesian classification, and decision tree were used to divide databases into molecules predicted to be active and those predicted to be inactive. Training and predicted activities were treated as binary. The database was generated for the ligands of five different biological targets which have been the object of intense drug discovery efforts: HIV-reverse transcriptase, COX2, dihydrofolate reductase, estrogen receptor, and thrombin. We report significant differences in the performance of the methods independent of the biological target and compound class. Different methods can have different applications; some provide particularly high enrichment, others are strong in retrieving the maximum number of actives. We also show that these methods do surprisingly well in predicting recently published ligands of a target on the basis of initial leads and that a combination of the results of different methods in certain cases can improve results compared to the most consistent method.

Algorithms↗
PubMed2003

Invasive front grading: reliability and usefulness in the management of oral squamous cell carcinoma.

BACKGROUND: The value of histological grading was examined with emphasis on reliability of assessment in 102 cases of intraoral squamous cell carcinoma from Northern Ireland with known outcome. METHODS: Two pathologists independently graded the invasive tumour front blinded to the stage and outcome. RESULTS: Intraobserver agreement was acceptable but interobserver agreement was not satisfactory. The degree of keratinisation was assessed most consistently while nuclear polymorphism was the least reliable feature. Multivariate survival analysis showed that the total grading score was associated with overall survival while the pattern of tumour invasion was the most valuable feature in estimating regional lymph node involvement. The number of positive lymph nodes was strongly associated with regional relapse, while the treatment modality and status of the surgical margins correlated with local relapse. CONCLUSIONS: Grading of selected features in OSCC is reliable and can facilitate treatment planning.

Adult↗
PubMed1977

Development of psychomotor objectives for classroom of clinical education in physical therapy.

A systematic method for developing and evaluating psychomotor objectives for students in clinical and laboratory education was presented. In preparing objectives for motor learning, the instructor must perform a task analysis of each skill to be taught. The analysis involves 1) identifying components of the movement, 2) salient features of facilitating stimuli, such as placement of grips, intensity and speed of stimulation, and 3) timing of commands or resistance. The senses, such as vision or kinesthesia, through which the student is to perceive the selected features must be chosen. After determining the level of skill to be attained, teaching methods are easily fit to the objectives. A simple checklist aids evaluation of student competency and provides immediate feedback on performance level.

Humans↗
PubMed2004

Order and disorder in bacterial genomes.

The availability of sequenced bacterial genomes allows a deeper understanding of their organizational features that are related with fundamental cellular processes such as coordinated gene expression, chromosome replication and cell division. Nevertheless, recent genome comparisons and experimental work highlighted the fluidity of bacterial chromosomes, including genome rearrangements that imperil the selective features of chromosome order. As a result, the clash between elements generating rearrangements and chromosome organization is a classic case of evolutionary conflict.

Bacteria↗
PubMed2006

A fully automated method for lung nodule detection from postero-anterior chest radiographs.

In the past decades, a great deal of research work has been devoted to the development of systems that could improve radiologists' accuracy in detecting lung nodules. Despite the great efforts, the problem is still open. In this paper, we present a fully automated system processing digital postero-anterior (PA) chest radiographs, that starts by producing an accurate segmentation of the lung field area. The segmented lung area includes even those parts of the lungs hidden behind the heart, the spine, and the diaphragm, which are usually excluded from the methods presented in the literature. This decision is motivated by the fact that lung nodules may be found also in these areas. The segmented area is processed with a simple multiscale method that enhances the visibility of the nodules, and an extraction scheme is then applied to select potential nodules. To reduce the high number of false positives extracted, cost-sensitive support vector machines (SVMs) are trained to recognize the true nodules. Different learning experiments were performed on two different data sets, created by means of feature selection, and employing Gaussian and polynomial SVMs trained with different parameters; the results are reported and compared. With the best SVM models, we obtain about 1.5 false positives per image (fp/image) when sensitivity is approximately equal to 0.71; this number increases to about 2.5 and 4 fp/image when sensitivity is = 0.78 and = 0.85, respectively. For the highest sensitivity (= 0.92 and 1.0), we get 7 or 8 fp/image.

Algorithms↗
PubMed2003

The use of artificial neural networks for the selection of the most appropriate formulation and processing variables in order to predict the in vitro dissolution of sustained release minitablets.

The objective of this work was to apply artificial neural networks (ANNs) to examine the relative importance of various factors, both formulation and process, governing the in-vitro dissolution from enteric-coated sustained release (SR) minitablets. Input feature selection (IFS) algorithms were used in order to give an estimate of the relative importance of the various formulation and processing variables in determining minitablet dissolution rate. Both forward and backward stepwise algorithms were used as well as genetic algorithms. Networks were subsequently trained using the back propagation algorithm in order to check whether or not the IFS process had correctly located any unimportant inputs. IFS gave consistent rankings for the importance of the various formulation and processing variables in determining the release of drug from minitablets. Consistent ranking was achieved for both indices of the release process; ie, the time taken for release to commence through the enteric coat (T(lag)) and that for the drug to diffuse through the SR matrix of the minitablet into the dissolution medium (T9(0-10)). In the case of the T(lag) phase, the main coating parameters, along with the original batch blend size and the blend time with lubricant, were found to have most influence. By contrast, with the T(90-10 phase), the amounts of matrix forming polymer and direct compression filler were most important. In the subsequent training of the ANNs, removal of inputs regarded as less important led to improved network performance. ANNs were capable of ranking the relative importance of the various formulations and processing variables that influenced the release rate of the drug from minitablets. This could be done for all main stages of the release process. Subsequent training of the ANN verified that removal of less relevant inputs from the training process led to an improved performance from the ANN.

Artificial Intelligence↗
PubMed2004

Feature reduction and morphological processing for hyperspectral image data.

An automatic target detection system that uses hyperspectral (HS) imagery is proposed. HS images contain both spatial and spectral response information that provides detailed descriptions of an object. These new, to our knowledge, sensor data are useful in automatic target recognition applications. To provide discrimination information from the HS images and to select features that generalize well, we describe a new, to our knowledge, high-dimensional generalized discriminant feature-extraction algorithm and compare its performance with that of other feature-reduction methods for two HS target detection applications (mine and vehicle detection) by using a nearest-neighbor classifier. We also advance an approach to simultaneously optimize both spatial and spectral responses.

Journal Article↗
PubMed1999

Temporal encoding of the voice onset time phonetic parameter by field potentials recorded directly from human auditory cortex.

Voice onset time (VOT) is an important parameter of speech that denotes the time interval between consonant onset and the onset of low-frequency periodicity generated by rhythmic vocal cord vibration. Voiced stop consonants (/b/, /g/, and /d/) in syllable initial position are characterized by short VOTs, whereas unvoiced stop consonants (/p/, /k/, and t/) contain prolonged VOTs. As the VOT is increased in incremental steps, perception rapidly changes from a voiced stop consonant to an unvoiced consonant at an interval of 20-40 ms. This abrupt change in consonant identification is an example of categorical speech perception and is a central feature of phonetic discrimination. This study tested the hypothesis that VOT is represented within auditory cortex by transient responses time-locked to consonant and voicing onset. Auditory evoked potentials (AEPs) elicited by stop consonant-vowel (CV) syllables were recorded directly from Heschl's gyrus, the planum temporale, and the superior temporal gyrus in three patients undergoing evaluation for surgical remediation of medically intractable epilepsy. Voiced CV syllables elicited a triphasic sequence of field potentials within Heschl's gyrus. AEPs evoked by unvoiced CV syllables contained additional response components time-locked to voicing onset. Syllables with a VOT of 40, 60, or 80 ms evoked components time-locked to consonant release and voicing onset. In contrast, the syllable with a VOT of 20 ms evoked a markedly diminished response to voicing onset and elicited an AEP very similar in morphology to that evoked by the syllable with a 0-ms VOT. Similar response features were observed in the AEPs evoked by click trains. In this case, there was a marked decrease in amplitude of the transient response to the second click in trains with interpulse intervals of 20-25 ms. Speech-evoked AEPs recorded from the posterior superior temporal gyrus lateral to Heschl's gyrus displayed comparable response features, whereas field potentials recorded from three locations in the planum temporale did not contain components time-locked to voicing onset. This study demonstrates that VOT at least partially is represented in primary and specific secondary auditory cortical fields by synchronized activity time-locked to consonant release and voicing onset. Furthermore, AEPs exhibit features that may facilitate categorical perception of stop consonants, and these response patterns appear to be based on temporal processing limitations within auditory cortex. Demonstrations of similar speech-evoked response patterns in animals support a role for these experimental models in clarifying selected features of speech encoding.

Auditory Cortex↗
PubMed2003

Discriminative motifs.

This paper takes a new view of motif discovery, addressing a common problem in existing motif finders. A motif is treated as a feature of the input promoter regions that leads to a good classifier between these promoters and a set of background promoters. This perspective allows us to adapt existing methods of feature selection, a well-studied topic in machine learning, to motif discovery. We develop a general algorithmic framework that can be specialized to work with a wide variety of motif models, including consensus models with degenerate symbols or mismatches, and composite motifs. A key feature of our algorithm is that it measures overrepresentation while maintaining information about the distribution of motif instances in individual promoters. The assessment of a motif's discriminative power is normalized against chance behaviour by a probabilistic analysis. We apply our framework to two popular motif models and are able to detect several known binding sites in sets of co-regulated genes in yeast.

Algorithms↗
PubMed2006

The use of artificial neural networks for the selection of the most appropriate thermal parameters and for the classification of a set of phenylcarbamic acid derivates.

The objective of this work was to apply artificial neural networks (ANNs) to the classification group of 43 derivatives of phenylcarbamic acid. To find the appropriate clusters Kohonen topological maps were employed. As input data, thermal parameters obtained during DSC and TG analysis were used. Input feature selection (IFS) algorithms were used in order to give an estimate of the relative importance of various input variables. Additionally, sensitivity analysis was carried out to eliminate less important thermal variables. As a result, one classification model was obtained, which can assign our compounds to an appropriate class. Because the classes contain groups of molecules structurally related, it is possible to predict the structure of the compounds (for example the position of the substitution alkoxy group in the phenyl ring) on the basis of obtained parameters.

Calorimetry, Differential Scanning↗
PubMed2004

An ERP index of task relevance evaluation of visual stimuli.

Recently several event-related potential attention studies have described a prefrontal positivity at about the same latency as the posterior N2 (approximately 200-300 ms), variously termed the frontal selection positivity (FSP), the anterior P2 (P2a), or the frontal P3 (P3f). These components have a similar spatio-temporal distribution and similar eliciting properties, suggesting that they represent the same component. However, these components have been differentially interpreted as arising from neural systems of feature selection, stimulus evaluation, or response production. The present study employed a visual target detection (oddball) design with different response conditions: passive (no response), overt (keypress), and covert (silent count), to examine the impact of task relevance and response production on the frontal P2a. The results showed that the P2a was present to task-relevant stimuli but had the same scalp topography and estimated source-dipole locations in both overt and covert responding, indicative of an index of stimulus evaluation, rather than response production.

Attention↗
PubMed2001

A polyphasic taxonomic study of thermophilic bacilli from shallow, marine vents.

Eighty-seven thermophilic, aerobic, spore-forming bacteria were isolated from shallow, marine, thermal vents of the Eolian Islands (Italy) and tested for a broad spectrum of phenotypic characteristics. A numerical taxonomy study was performed on these isolates and 8 thermophilic Bacillus and Geobacillus reference strains by 89 selected features. Results from cluster analysis showed the formation of nine clusters. Most of the isolates (83%) fell into several phenetically well distinguished clusters, loosely related to Geobacillus thermodenitrificans. The remaining isolates grouped together with different reference strains. Eighteen isolates, representative of the different clusters, were selected for subsequent genotypic characterisation, including partial 16S rDNA sequence analysis of 18 strains and almost complete 16S rDNA sequences of 9 strains. Subsequent DNA/DNA reassociation studies and determination of the base composition of DNA identified seven isolates as Geobacillus thermodenitrificans, two isolates as G. thermoleovorans and one isolate as Bacillus pallidus. Four isolates represented two novel species of Bacillus. The remaining four represented novel Geobacillus species, one of which has recently been described as Bacillus vulcani DSMZ 13174 T.

Bacillus↗
PubMed1986

Classification of lung carcinoma by means of digital nuclear image analysis.

An investigation was performed of the maximum discriminating efficiency for each subgroup of digital nuclear image features and of the overall classification of nuclei from three types of human lung carcinomas in histologic sections: adenocarcinoma, small-cell carcinoma and squamous-cell carcinoma. The results indicate that, for each subgroup of features, the nuclei of the small-cell carcinomas are generally "correctly" classified in a higher percentage (80% to 100%) than are the nuclei of the adenocarcinomas (46% to 74%) and squamous-cell carcinomas (29% to 68%). The discriminant analysis for the overall classification selected features from most of the subgroups, suggesting that it is useful to perform nuclear image analysis with many subgroups having different properties. The overall classifications for the nuclei of the adenocarcinomas, small-cell carcinomas and squamous-cell carcinomas were, respectively, 81.4%, 93.2% and 74.7%. Before this technique can be applied to histopathologic diagnosis, a larger number of unselected lung carcinomas must be evaluated.

Adenocarcinoma↗
PubMed2006

Selective stabilization and synaptic specificity: a new cell-biological model.

How are appropriate connections between neurons sorted from the overwhelming surplus of potential, yet inappropriate, connections? Despite the apparently improbable nature of the process, brains wire themselves with a high degree of reproducibility that has been conserved across evolutionary history. Here, we outline a viable cell-biological model for generating synaptic specificity that features selection of nascent synapses based on adhesion and recognition. This process uses the highly dynamic and stochastic nature of intracellular trafficking to generate reproducible patterns of synaptic connectivity.

Animals↗
PubMedDate not supplied

Conformational and physicochemical DNA features specific for transcription factor binding sites.

MOTIVATION: A reliable recognition of transcription factor binding sites is essential for analysis of regulatory genomic sequences. The experimental data make evident an important role of DNA conformational features for site functioning. However, Internet-available tools for revealing conformational and physicochemical DNA features significant for the site functioning and subsequent use of these features for site recognition have not been developed up to now. RESULTS: We suggest an approach for revealing significant conformational and physicochemical properties of functional sites implemented in the database B-DNA-VIDEO. This database is designed to study the sets of various transcription factor binding sites, providing evidence that transcription factor binding sites are characterized by specific sets of significant conformational and physicochemical DNA properties. For a fixed site, by using the B-DNA features selected for this site recognition, the C-program recognizing this site may be generated, control tested and stored in the database B-DNA-VIDEO. Each B-DNA-VIDEO entry links to the Web-applet recognizing the site, whose significant B-DNA features are stored in this entry as the 'site recognition programs'. The pairwise linked entry-applet pairs are compiled within the B-DNA-VIDEO system, which is simultaneously the database and the program tools package applicable immediately for recognizing the sites stored in the database. Indeed, this is the novelty. Hence, B-DNA-VIDEO is the Web resource of both 'searching for static data' and 'active computation' type, that is why it was called an 'activated database'. AVAILABILITY: B-DNA-VIDEO is available at http://wwwmgs.bionet.nsc.ru/systems/BDNAVideo/ and the mirror site at http://www.cbil.upenn.edu/mgs/systems/c onsfreq/.

Base Sequence↗
PubMed2006

Spatiotemporal receptive fields of peripheral afferents and cortical area 3b and 1 neurons in the primate somatosensory system.

Neurons in area 3b have been previously characterized using linear spatial receptive fields with spatially separated excitatory and inhibitory regions. Here, we expand on this work by examining the relationship between excitation and inhibition along both spatial and temporal dimensions and comparing these properties across anatomical areas. To that end, we characterized the spatiotemporal receptive fields (STRFs) of 32 slowly adapting type 1 (SA1) and 21 rapidly adapting peripheral afferents and of 138 neurons in cortical areas 3b and 1 using identical random probe stimuli. STRFs of peripheral afferents consist of a rapidly appearing excitatory region followed by an in-field (replacing) inhibitory region. STRFs of SA1 afferents also exhibit flanking (surround) inhibition that can be attributed to skin mechanics. Cortical STRFs had longer time courses and greater inhibition compared with peripheral afferent STRFs, with less replacing inhibition in area 1 neurons compared with area 3b neurons. The greater inhibition observed in cortical STRFs point to the existence of underlying intracortical mechanisms. In addition, the shapes of excitatory and inhibitory lobes of both peripheral and cortical STRFs remained mostly stable over time, suggesting that their feature selectivity remains constant throughout the time course of the neural response. Finally, the gradual increase in the proportion of surround inhibition from the periphery to area 3b to area 1, and the concomitant decrease in response linearity of these neurons indicate the emergence of increasingly feature-specific response properties along the somatosensory pathway.

Afferent Pathways↗
PubMed1995

Nitric oxide-based possibilities for pharmacotherapy.

The goal of nitric oxide (NO) based pharmacotherapy is to reach proper homeostasis of NO metabolism in the target tissue where endogenous production of NO is either too weak or excessively increased. In addition to the classic NO-based therapy of cardiovascular conditions with nitrates, a variety of new therapeutic possibilities have emerged including sexual disorders, gastrointestinal system, immunology, tumour growth regulation and respiratory disorders. NO levels of target tissues can be affected directly by NO donors, or indirectly by increasing the level of L-arginine, a substrate of nitric oxide synthase (NOS). While increased production of NO by induceable NO (iNOS) by, for example, cytokines does not at present seem therapeutically meaningful, increased NO production by constitutive NOS (cNOS) may be involved in the beneficial effects of ACE inhibitors or oestrogens. NO production may be pharmacologically decreased by inhibition of expression of iNOS by glucocorticoids while both cNOS and iNOS derived NO production is inhibited by administration of false substrates, for example L-NAME. Additionally, the respiratory system and related vessels can be reached directly and more selectively by inhalation of pure NO gas. Possible problems in administering NO and perhaps some NO-donors include the toxic nature of the compound itself whereby vital enzyme systems may be inhibited and tissue damaging radicals formed. Future prospects of NO-based pharmacotherapy may feature selective ligands to different NOS isoforms and tissue selective donors that release NO in a controlled fashion.

Adult↗
PubMed2005

Classification of structural images via high-dimensional image warping, robust feature extraction, and SVM.

This paper presents a method for classification of medical images, using machine learning and deformation-based morphometry. A morphological representation of the anatomy of interest is first obtained using high-dimensional template warping, from which regions that display strong correlations between morphological measurements and the classification (clinical) variable are extracted using a watershed segmentation, taking into account the regional smoothness of the correlation map which is estimated by a cross-validation strategy in order to achieve robustness to outliers. A Support Vector Machine-Recursive Feature Elimination (SVM-RFE) technique is then used to rank computed features from the extracted regions, according to their effect on the leave-one-out error bound. Finally, SVM classification is applied using the best set of features, and it is tested using leave-one-out. The results from a group of 61 brain images of female normal controls and schizophrenia patients demonstrate not only high classification accuracy (91.8%) and steep ROC curves, but also exceptional stability with respect to the number of selected features and the SVM kernel size.

Algorithms↗