SEARCH · PubMed Health

Results for “Model performance”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 955 records · Page 53Linked to original sources

PubMed2005

Impact of comorbid panic and posttraumatic stress disorder on outcomes of collaborative care for late-life depression in primary care.

OBJECTIVE: Comorbid anxiety disorders may result in worse depression treatment outcomes. The authors evaluated the effect of comorbid panic disorder and posttraumatic stress disorder (PTSD) on response to a collaborative-care intervention for late-life depression in primary care. METHODS: A total of 1,801 older adults with depression were randomized to a collaborative-care depression treatment model versus usual care and assessed at baseline, 3, 6, and 12 months, comparing differences among participants with comorbid panic disorder (N=262) and PTSD (N=191) and those without such comorbid anxiety disorders. RESULTS: At baseline, patients with comorbid anxiety reported higher levels of psychiatric and medical illness, greater functional impairment, and lower quality of life. Participants without comorbid anxiety who received collaborative care had early and lasting improvements in depression compared with those in usual care. Participants with comorbid panic disorder showed similar outcomes, whereas those with comorbid PTSD showed a more delayed response, requiring 12 months of intervention to show a significant effect. At 12 months, however, outcomes were comparable. Interactions of intervention status by comorbid PTSD or panic disorder were not statistically significant, suggesting that the collaborative-care model performed significantly better than usual care in depressed older adults both with and without comorbid anxiety. CONCLUSIONS: Collaborative care is more effective than usual care for depressed older adults with and without comorbid panic disorder and PTSD, although a sustained treatment response was slower to emerge for participants with PTSD. Intensive and prolonged follow-up may be needed for depressed older adults with comorbid PTSD.

Age Factors↗

PubMed2004

Orthoses as enablers of occupation: client-centred splinting for better outcomes.

BACKGROUND: Orthotic intervention (splinting) may have become an end unto itself in the minds of therapists and clients rather than the means to enable optimal occupational performance. Some policy makers and payers seem to hold the belief that orthoses/splints are mere technical aids and as such do not require professional skill and expertise. NARRATIVES: Three client stories demonstrate how iterative collaboration and follow-up help achieve client-identified objectives. DISCUSSION: Client input is an important component of the process and an orthosis must fit into the person's lifestyle, especially if required for long-term use. Six essential considerations when providing orthoses to meet occupational goals are emphasized: client-centredness, comfort, cosmesis, convenience, less is more and follow-up. Use of the Canadian Occupational Performance Model for intervention planning and as an outcome measure is demonstrated. PRACTICE IMPLICATIONS: Orthoses that are thoughtfully designed with client input and carefully constructed can make a difference in a person's life by relieving pain, providing protection and joint stabilization and enabling valued occupations.

Adult↗

PubMed1992

Visibility of transmissive liquid crystal displays under dynamic lighting conditions.

Liquid crystal displays maintain superior contrast under bright sunlight conditions compared with cathode-ray tube displays. This attribute, along with reduced weight, volume, and power requirements, make liquid crystal displays especially desirable for use in vehicular applications. The present experiment was designed to determine the luminance requirements for transmissive liquid crystal displays viewed under a wide range of lighting conditions typical of many vehicular environments. Both the ambient illumination incident on the display surface and the forward-field-of-view adaptation luminance were parametrically manipulated. The visual task involved speeded spatial discrimination responses for detection of the gap in a modified Landolt-C acuity target. The results indicate that under the worst lighting conditions tested, a display luminance of approximately 180 cd/m2 yielded asymptotic spatial discrimination performance. The results have been incorporated into a predictive visual performance model for spatial discrimination tasks, which describes the present data well and represents a first step toward a device-independent model of display visibility.

Adult↗

PubMed1984

A calculator program for individualizing factor VIII dosage.

We developed a calculator program to individualize Factor VIII dosage on the basis of the concentration-time data obtained after a single test-dose. The program is designed for the Hewlett-Packard 41-CV calculator. The calculation procedure is the 1977 version of the Sawchuk and Zaske method. The program performs model parameter estimation through a nonlinear iterative least-squares technique (the modified Gauss-Newton method). We tested our program in three patients with classic hemophilia who required multiple-dose treatment with Factor VIII. In each patient, individual kinetic parameters were estimated from the serial plasma levels measured after the test-dose. The predicted concentration-time curve resulting from all the administered doses was calculated based on the estimated kinetic parameters. Good agreement between predicted and measured levels during the multiple-dose regimen was observed in all patients.

Computers↗

PubMed2005

Modelling the cost implications of using carboxymethylcellulose dressing compared with gauze in the management of surgical wounds healing by secondary intention in the US and UK.

OBJECTIVE: To estimate the costs of using carboxymethyl cellulose dressing (CMCD; Aquacel* Hydrofiber) compared to gauze in managing surgical wounds healing by secondary intention in the US and UK. STUDY DESIGN: This was a modelling study performed from the perspective of payers (i.e. the hospital and community sector in the US and the National Health Service (NHS) in the UK). METHODS: Clinical outcomes attributable to managing surgical wounds healing by secondary intention with gauze were obtained from the published literature in the English language. There were no published studies on wounds healing by secondary intention with CMCD. Hence, the analysis conservatively assumed that wound healing rates associated with gauze would be the same for CMCD. These data were combined with resource utilisation estimates derived from a panel of clinicians enabling us to perform decision modelling. The models were used to determine the expected direct healthcare costs eight weeks after the surgical wounds were dressed by CMCD or gauze and left to heal by secondary intention in the US and UK. RESULTS: All wounds are expected to heal within eight weeks, irrespective of dressing. Managing abscesses and other surgical wounds with CMCD instead of gauze in the US is expected to reduce costs by 4% in both wound types (i.e. $247 and $507 respectively) per patient over eight weeks. In the UK, managing abscesses and other surgical wounds with CMCD instead of gauze is expected to reduce costs by 30% (574 pounds) and 12% (581 pounds) respectively per patient over eight weeks. The lower cost of managing CMCD-treated patients is due to decreased nursing costs associated with a lower frequency of CMCD changes compared to gauze dressing changes. CONCLUSION: Dressing surgical wounds healing by secondary intention with CMCD instead of gauze is expected to lead to a reduction in healthcare costs in both the US and UK. Hence, the purchase price of a dressing is not indicative of the cost effectiveness of a given method of surgical wound care.

Bandages, Hydrocolloid↗

PubMed2005

A prediction rule to identify low-risk patients with heart failure.

OBJECTIVES: To derive a prediction rule using data available in the emergency department (ED) to identify a group of patients hospitalized for the treatment of heart failure who are at low risk of death and serious complications. METHODS: The authors analyzed data for all 33,533 patients with a primary hospital discharge diagnosis of heart failure in 1999 who were admitted from EDs in Pennsylvania. Candidate predictors were demographic and medical history variables and the most abnormal examination or diagnostic test values measured in the ED (vital signs only) or on the first day of hospitalization. The authors constructed classification trees to identify a subgroup of patients with an observed rate of death or serious medical complications before discharge < 2%; the tree that identified the subgroup with the lowest rate of this outcome and an inpatient mortality rate < 1% was chosen. RESULTS: Within the entire cohort, 4.5% of patients died and 6.8% survived to hospital discharge after experiencing a serious medical complication. The prediction rule used 21 prognostic factors to classify 17.2% of patients as low risk; 19 (0.3%) died and 59 (1.0%) survived to hospital discharge after experiencing a serious medical complication. CONCLUSIONS: This clinical prediction rule identified a group of patients hospitalized from the ED for the treatment of heart failure who were at low risk of adverse inpatient outcomes. Model performance needs to be examined in a cohort of patients with an ED diagnosis of heart failure and treated as outpatients or hospitalized.

Aged↗

PubMed1989

Increased risk of lung cancer, non-Hodgkin's lymphoma, and leukemia following Hodgkin's disease.

The risk of second cancers (SCs) was assessed in 744 patients with Hodgkin's disease (HD) admitted to The Netherlands Cancer Institute from 1966 to 1983. Sixty-nine SCs were observed one month or more after start of first treatment. These included 14 cases of lung cancer, nine cases of non-Hodgkin's lymphoma (NHL), 16 cases of leukemia, and six cases of the myelodysplastic syndrome (MDS). The median interval between the diagnosis of HD and that of second lung cancer, NHL, and leukemia was 8.1, 13.3, and 5.7 years, respectively. The overall relative risks (RR) (observed/expected [O/E] ratios) of developing lung cancer, NHL, and leukemia were 4.9 (95% confidence limit [CL], 2.7 to 8.2), 31.0 (95% CL, 14.2 to 58.9) and 45.7 (95% CL, 26.1 to 74.2), respectively. At 15 years the cumulative risk of developing an SC amounted to 20.6% +/- 2.9%. The 15-year estimates of lung cancer, NHL, and leukemia were 6.2% +/- 1.9%, 5.9% +/- 2.1% and 6.3% +/- 1.7%, respectively. Increased lung cancer risk following HD has not frequently been clearly demonstrated before; that we were able to demonstrate such risk may be due to the completeness of follow-up over long periods that could be achieved in this study. Excess lung cancer risk was only noted in treatment regimens with radiotherapy (RT); also, all lung cancers arose in irradiation fields. Excess risk of leukemia was only found in treatment regimens involving chemotherapy (CT). For NHL, combined modality treatment was shown to be the most important risk factor. Risk of lung cancer and NHL increased with time since diagnosis. A time-dependent covariate analysis (Cox model) performed on leukemia and MDS showed an increasing risk with intensity of CT, age (greater than 40 years), and a splenectomy.

Adult↗

PubMed1991

Prognostic factors in unselected patients with nonseminomatous metastatic testicular cancer: a multicenter experience.

Between 1981 and 1986, 200 consecutive patients with metastatic nonseminomatous testicular cancer were entered into the Swedish Norwegian Testicular Cancer (SWENOTECA) project from 14 hospitals. The treatment plan was four chemotherapy cycles (cisplatin, vinblastine, and bleomycin) followed by surgical resection of residual tumor masses. After a median observation time of 75 months, the overall 5-year survival rate was 82%. In a univariate analysis, the following parameters influenced the prognosis significantly: the extent of the disease (Medical Research Council [MRC] grouping); the prechemotherapy levels of serum alpha-fetoprotein (AFP), human chorionic gonadotropin (HCG), and lactate dehydrogenase (LDH); the patients' age; the presence of extrapulmonary hematogeneous metastases; and/or particularly large lymph node metastases. Patients fared better when more than 3 weeks elapsed between orchiectomy and start of chemotherapy as compared with those who were treated within this interval. The place of treatment (a large oncology unit v smaller units) also represented a significant prognostic factor for patients with large-volume (LV) and very-large-volume (VLV) disease combined. Multivariate analysis (Cox regression proportional hazards model) performed in all 193 assessable patients showed the following adverse prognostic factors: high-volume metastatic burden, age older than 35 years, prechemotherapy AFP greater than 500 micrograms/L and/or HCG greater than 1,000 U/L, and an interval between orchiectomy and start of chemotherapy of less than 3 weeks. The place of treatment also significantly influenced the final outcome. If patients with LV and VLV disease were combined, the presence of two of the following risk factors represented an additional prognostic factor: AFP greater than 1,000 micrograms/L, HCG greater than 10,000 U/L, liver metastases, brain metastases, bone metastases, retroperitoneal tumor greater than or equal to 10 cm, and mediastinal tumor greater than or equal to 5 cm.

Adolescent↗

PubMed2005

Use of an EMG-driven biomechanical model to study virtual injuries.

INTRODUCTION: How the CNS activates muscles to produce coordinated movement is a matter of debate and great interest. We are attempting to answer this question, in part, by investigating how individual muscles and groups of muscles are activated under different physiologic and environmental conditions. We have developed an EMG-driven virtual arm to assist in this endeavor. PURPOSE: To demonstrate how the virtual arm can be used to simulate a neuromuscular injury and to examine whether a virtual injury can evoke a change in muscle activation patterns. METHODS: The virtual arm is a three-dimensional graphical representation and biomechanical model of a human arm including the major flexor and extensor muscles crossing the elbow. The muscles are actuated based on experimentally recorded electromyograms. A Hill-type muscle model was used to predict muscle forces, which in turn were used to move the graphical display of the arm on the screen. Two subjects, one considered highly trained and the other a novice, participated in this study. Virtual movements, before and after simulating an injury were evaluated, and model performance was assessed by comparing the virtual arm-predicted moment and the actual moment generated by the subjects. RESULTS: The highly trained subject was proficient at controlling the virtual arm. For this subject, simulating a neuromuscular injury evoked a different pattern of activation compared to the preinjured state. CONCLUSIONS: The virtual arm may be a useful tool for the study of motor coordination and how muscle activation patterns change in response to injury. Future work involving more subjects and experimental conditions is planned to better assess the efficacy of the virtual arm as a research tool for investigating motor control strategies.

Arm↗

PubMed2003

Simultaneous measurements of particle backscattering and extinction coefficients and wind velocity by lidar with a Mach-Zehnder interferometer: principle of operation and performance assessment.

The development of remote-sensing instruments that can be used to monitor several parameters at the same time is important for the study of complex processes such as those that control climate and environment. In this paper the performance of a new concept of lidar receiver that allows for the direct measurement of aerosol and cloud optical properties simultaneously with wind velocity is investigated. This receiver uses a Mach-Zehnder interferometer. Two different configurations, either with four photometric output channels or with fringe imaging on a multichannel detector, are studied. Analytical expressions of the statistical errors are given under the assumption of Gaussian signal spectra. It is shown that similar accuracies can be achieved for both configurations. Performance modeling of the retrieval of semitransparent cloud optical scattering properties and wind velocity was done at different operation wavelengths for a Nd:YAG laser source. Results for such a lidar system onboard an aircraft flying at an altitude of 12 km show that for semitransparent clouds the best results were obtained at 355 nm, with relative standard deviations of 0.5% and 5% for the backscatter and extinction coefficients, respectively, together with a velocity accuracy of 0.2 ms(-1). The accuracy of optical properties retrieved for boundary layer aerosols are comparable, whereas the velocity accuracy is decreased to 1 ms(-1). Finally, an extrapolation to a large 355-nm spaceborne lidar shows accuracies in the range from 2.5% to 5% for the backscatter coefficient and from 10% to 15% for the extinction coefficient together with a vertical wind speed accuracy of better than 0.5 ms(-1) for semitransparent clouds and boundary layer, with a vertical resolution of 500 m and a 100 shot averaging.

Journal Article↗

PubMed2006

Principal component-based radiative transfer model for hyperspectral sensors: theoretical concept.

Modern infrared satellite sensors such as the Atmospheric Infrared Sounder (AIRS), the Cross-Track Infrared Sounder (CrIS), the Tropospheric Emission Spectrometer (TES), the Geosynchronous Imaging Fourier Transform Spectrometer (GIFTS), and the Infrared Atmospheric Sounding Interferometer (IASI) are capable of providing high spatial and spectral resolution infrared spectra. To fully exploit the vast amount of spectral information from these instruments, superfast radiative transfer models are needed. We present a novel radiative transfer model based on principal component analysis. Instead of predicting channel radiance or transmittance spectra directly, the principal component-based radiative transfer model (PCRTM) predicts the principal component (PC) scores of these quantities. This prediction ability leads to significant savings in computational time. The parameterization of the PCRTM model is derived from the properties of PC scores and instrument line-shape functions. The PCRTM is accurate and flexible. Because of its high speed and compressed spectral information format, it has great potential for superfast one-dimensional physical retrieval and for numerical weather prediction large volume radiance data assimilation applications. The model has been successfully developed for the NAST-I and AIRS instruments. The PCRTM model performs monochromatic radiative transfer calculations and is able to include multiple scattering calculations to account for clouds and aerosols.

Journal Article↗

PubMed2005

Usability of touch-panel interfaces for older adults.

The usability of a touch-panel interface was compared among young, middle-aged, and older adults. In addition, a performance model of a touch panel was developed so that pointing time could be predicted with higher accuracy. Moreover, the target location to which a participant could point most quickly was determined. The pointing time with a PC mouse was longer for the older adults than for the other age groups, whereas there were no significant differences in pointing time among the three age groups when a touch-panel interface was used. Pointing to the center of a square target led to the fastest pointing time among nine target locations. Based on these results, we offer some guidelines for the design of touch-panel interfaces and show implications for users of different age groups. Actual or potential applications of this research include designing touch-panel interfaces to make them accessible for older adults and predicting movement times when users operate such devices.

Adult↗

PubMed2001

Systemic neutrophil response resulting from dental plaque accumulation.

BACKGROUND: There is considerable current interest in putative relationships between oral and systemic diseases. Since the host response to oral bacteria may be the critical link in this association, our hypothesis was that dental plaque accumulation in healthy subjects would elicit a systemic inflammatory response. METHODS: Twenty-three healthy subjects, aged 18 to 25, participated in a 4-phase study. An initial hygiene phase was followed by a 21-day experimental phase (the so-called experimental gingivitis model) in which subjects refrained from all oral hygiene practices, thus permitting the accumulation of bacterial plaque. At days 0, 7, and 21 total and differential peripheral white blood cell (wbc) counts, together with full mouth plaque and gingivitis scores, were recorded. Following a 28-day recovery phase, in which normal oral hygiene practices were resumed, subjects entered the final 21-day control phase which mirrored the experimental phase but with subjects maintaining normal oral hygiene practices. RESULTS: The experimental model performed as anticipated with a correlation between plaque and gingivitis scores of 0.95, also reflecting subject compliance. Total wbc and neutrophil counts increased during the experimental phase. Furthermore, comparison of neutrophil counts between the experimental and control phases demonstrated a significantly higher cell count for the experimental phase on both days 7 and 21 (P= 0.0301 and 0.009, respectively). For total wbc, this was significant on day 21 (P= 0.0262). CONCLUSION: The results of this study support the hypothesis that the accumulation of dental plaque can result in a measurable systemic inflammatory response, providing further in vivo data to support a mechanistic relationship between oral and systemic pathology.

Adolescent↗

PubMed2026

A machine learning-derived intratumoral heterogeneity-related signature predicts the prognosis for and therapeutic response in patients with skin cutaneous melanoma.

BACKGROUND: Reliable biomarkers for predicting prognosis and therapeutic response in skin cutaneous melanoma (SKCM) remain limited. This study aimed to develop an intratumoral heterogeneity (ITH)-related prognostic signature for SKCM using integrative machine learning. METHODS: RNA sequencing (RNA-seq) data from 472 SKCM patients in The Cancer Genome Atlas (TCGA) and 214 patients in the GSE65904 cohort were analyzed. ITH scores were calculated using the DEPTH2 algorithm. Differentially expressed genes (DEGs) were identified between high- and low-ITH groups [|log2fold change (FC)| ≥1, false discovery rate (FDR) <0.05]. Based on 38 prognostic DEGs identified by univariate Cox regression, we employed an integrative framework of 101 machine learning algorithm combinations to construct prognostic models in the TCGA training cohort. The model with the highest average concordance index (C-index) was validated in the GSE65904 cohort and selected as the prognostic ITH-related signature (PIRS). Associations of the PIRS risk score with tumor mutational burden (TMB), immune cell infiltration, immune checkpoint gene expression, and drug sensitivity were systematically evaluated. Model performance was assessed using receiver operating characteristic (ROC) curves and Cox regression analyses. RESULTS: A 38-gene PIRS was constructed using the plsRcox algorithm. Patients with high PIRS risk scores exhibited significantly poorer overall survival (OS) in both the TCGA and Gene Expression Omnibus (GEO) cohorts. The PIRS was identified as an independent prognostic factor, with area under the curve (AUC) values of 0.779, 0.734, and 0.756 for 1-, 3-, and 5-year survival, respectively. High-risk samples displayed significantly lower TMB (P<0.05), reduced immune and stromal cell infiltration (P<0.001), downregulated immune function, and decreased expression of immune checkpoint genes. Additionally, high- and low-PIRS risk score groups exhibited distinct sensitivity patterns to different classes of targeted agents. CONCLUSIONS: The machine learning-derived PIRS robustly predicts prognosis in SKCM patients. Its clinical application is promising for optimizing patient risk stratification and treatment decisions, though further prospective validation is warranted.

Skin cutaneous melanoma (SKCM)↗

PubMed2001

Management of childhood illness at health facilities in Benin: problems and their causes.

OBJECTIVES: To prepare for the implementation of Integrated Management of Childhood Illness (IMCI) in Benin, we studied the management of ill children younger than 5 years at outpatient health facilities. METHODS: We observed a representative sample of consultations; after each consultation, we interviewed caregivers and reexamined children. Health workers' performance was evaluated against IMCI guidelines. To identify determinants of performance, statistical modeling was performed and 6 focus groups with health workers were conducted to solicit their opinions. RESULTS: Altogether, 584 children were enrolled and 101 health workers were observed; 130 health workers participated in focus group discussions. Many serious deficiencies were found: incomplete assessment of children's signs and symptoms, incorrect diagnosis and treatment of potentially life-threatening illnesses, inappropriate prescription of dangerous sedatives, missed opportunities to vaccinate, and failure to refer severely ill children for hospitalization. Quantitative and qualitative analyses showed various health facility-, health worker-, caregiver-, and child-related factors as possible determinants of health worker performance. CONCLUSIONS: Action is urgently needed. Our results suggest that to improve health care delivery, interventions should target both the health system and the community level.

Benin↗

PubMed1998

Docetaxel. A pharmacoeconomic review of its use in the treatment of metastatic breast cancer.

UNLABELLED: Advanced breast cancer has a poor prognosis (median duration of survival about 2 years); thus, treatment options are largely palliative. Recent studies with the taxoids docetaxel and paclitaxel suggest that these agents are effective second-line therapy in patients with metastatic breast cancer. Objective response rates range from 30 to 57% with docetaxel monotherapy in clinical trials in patients with advanced disease who had failed to respond to previous therapy, and the median time to disease progression ranges from 17 to 20 weeks. Currently available pharmacoeconomic data relating to docetaxel in the treatment of women with metastatic breast cancer include a cost-minimisation study and 3 cost-utility studies which relate to the UK, French and US healthcare systems. All analyses compared docetaxel with paclitaxel. Since no direct clinical comparisons of these 2 agents have been performed, model assumptions were based on available, rather than comparative, clinical data. The French model also considered vinorelbine monotherapy as a comparator. Considering direct costs only, costs associated with docetaxel ranged from < 0.3% less to 13% more than those of paclitaxel. When utilities were evaluated, docetaxel was associated with incremental gains versus paclitaxel of 33 and 75 additional days of quality-adjusted health in the UK and US studies, respectively, and 22 additional days of quality-adjusted progression-free survival (QPFS) in the French analysis. Thus, the incremental cost utility of docetaxel versus paclitaxel was an estimated Pounds 2431 per quality-adjusted life-year (QALY) in the UK analysis (1994 values) and US$8615 per QALY in the US analysis (1997 values). The French model found that docetaxel cost 700 French francs less for the extra days of QPFS (1993 costs). CONCLUSIONS: Cost-utility analyses, compiled in the absence of direct comparative data, show that docetaxel offers utility gains versus paclitaxel in the treatment of metastatic breast cancer. The incremental cost for these gains is within the accepted range for healthcare interventions.

Antineoplastic Agents, Phytogenic↗

PubMed2002

Dynamic QSAR techniques: applications in drug design and toxicology.

The basic principles of 3-D quantitative structure-activity relationships (QSARs) analysis are discussed in the light of the fuzzy logic concept. According to that concept, the traditionally one chemical - one structure - one parameter value relationship in QSAR is suggested to be modified into one chemical finite set of structures - range of parameter values principle. In this respect, two recently developed techniques accounting for conformational flexibility in 3-D QSARs are reviewed. A basic assumption underlying both methods is that chemical behavior in complex biological systems is context-dependent. A molecule can exist and interact in a variety of conformations depending on the specificity of the endpoint under investigation and reaction media. It was demonstrated that selection of active conformer(s) in QSAR studies is a task as important as the selection of relevant molecular parameters. Specifically selected active conformers, rather than the lowest-energy states of the chemicals are suggested to be used in the correlative QSARs. The method for recognition the common reactivity pattern (COREPA) of structurally heterogeneous compounds that elicit similar biological behaviour is based on all energetically reasonable conformers of chemicals. The principle assumption of the method is that biologically similar chemicals should possess a commonality in their stereoelectronic (reactivity) pattern. Originally developed algorithms for conformer generation are presented in association with the QSAR methods accounting for conformational flexibility of chemicals. Applicability of the QSAR technique for selection active conformers is illustrated by presenting QSAR models derived for Ah binding affinity of PCBs and antimicrobial activity of rifamicin derivatives. Models for predicting estrogenic activity of structurally diverse chemicals and ACE inhibition exemplified the applicability of the COREPA method. The model performance is analyzed by the 3D screening exercise of large chemical inventories with subsequent experimental validation within the EDAEP project. Besides the impact of conformational flexibility of chemicals in 3D QSAR the role of different molecular descriptors is discussed with respect to their ability to describe molecular interactions with different specificity.

Combinatorial Chemistry Techniques↗

PubMed2004

Hierarchical Bayesian neural network for gene expression temporal patterns.

There are several important issues to be addressed for gene expression temporal patterns' analysis: first, the correlation structure of multidimensional temporal data; second, the numerous sources of variations with existing high level noise; and last, gene expression mostly involves heterogeneous multiple dynamic patterns. We propose a Hierarchical Bayesian Neural Network model to account for the input correlations of time course gene array data. The variations in absolute gene expression levels and the noise can be estimated with the hierarchical Bayesian setting. The network parameters and the hyperparameters were simultaneously optimized with Monte Carlo Markov Chain simulation. Results show that the proposed model and algorithm can well capture the dynamic feature of gene expression temporal patterns despite the high noise levels, the highly correlated inputs, the overwhelming interactions, and other complex features typically present in microarray data. We test and demonstrate the proposed models with yeast cell cycle temporal data sets. The model performance of Hierarchical Bayesian Neural Network was compared to other popular machine learning methods such as Nearest Neighbor, Support Vector Machine, and Self Organized Map.

Journal Article↗