SEARCH · PubMed Health

Results for “bootstrap”

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 595 records · Page 33Linked to original sources

PubMed2006

Data analysis of kinetic modelling used in drug stability studies: isothermal versus nonisothermal assays.

PURPOSE: Kinetic modelling was applied to predict the stability of cholecystokinin fragment CCK-4 in aqueous solution, which was analyzed by isothermal and nonisothermal methods using a validated stability indicating HPLC method. METHODS: The isothermal studies were performed in the temperature range 40 to 80 degrees C at pH 12 and ionic strength 0.01 M as constants, whereas nonisothermal stability studies were performed using a linear increasing temperature program, heating rate 0.25 degrees C/h and a temperature interval 40-82 degrees C. The isothermal studies require two-step linear regression to estimate the parameters, resulting in a well-defined confidence interval. Nonisothermal kinetic studies require nonlinear or linear regression by previous transformation of data to estimate the parameters. In this case, the two most popular approaches, derivative and integral, were used and compared. RESULTS: Under isothermal conditions, an apparent first-order degradation process was observed at all temperatures. The linear Arrhenius plot suggested that the CCK-4 degradation mechanism was the same within the studied temperature range, with quite large uncertainties due to the small number of degrees of freedom based only on the scatter in the plot, and giving an estimated shelf life at 25 degrees C of 35.2 days. The derivative approach yields high variability in the Arrhenius parameters, since they are dependent on the number of polynomial terms chosen, so several statistical criteria were applied to select the best model. The integral approach allows activation parameters to be calculated directly from experimental data, and provides results in good agreement with those of the traditional method, but have the advantage that the uncertainty in the final result directly reflects the goodness of fit of the experimental data to the chosen kinetic model. The application of the bootstrap technique to estimating confidence limits for the Arrhenius parameters and shelf life is also illustrated, and shows there is no difference between the asymptotic and bootstrap confidence intervals. CONCLUSIONS: Nonisothermal studies give us fast and valuable information about drug stability, although their potential for predicting isothermal behaviour is conditioned by the data analysis method applied.

Drug Stability↗

PubMed2006

Phylogenetic relationships of Steinernema Travassos, 1927 (Nematoda: Cephalobina: Steinernematidae) based on nuclear, mitochondrial and morphological data.

Entomopathogenic nematodes of the genus Steinernema are lethal parasites of insects that are used as biological control agents of several lepidopteran, dipteran and coleopteran pests. Phylogenetic relationships among 25 Steinernema species were estimated using nucleotide sequences from three genes and 22 morphological characters. Parsimony analysis of 28S (LSU) sequences yielded a well-resolved phylogenetic hypothesis with reliable bootstrap support for 13 clades. Parsimony analysis of mitochondrial DNA sequences (12S rDNA and cox 1 genes) yielded phylogenetic trees with a lower consistency index than for LSU sequences, and with fewer reliably supported clades. Combined phylogenetic analysis of the 3-gene dataset by parsimony and Bayesian methods yielded well-resolved and highly similar trees. Bayesian posterior probabilities were high for most clades; bootstrap (parsimony) support was reliable for approximately half of the internal nodes. Parsimony analysis of the morphological dataset yielded a poorly resolved tree, whereas total evidence analysis (molecular plus morphological data) yielded a phylogenetic hypothesis consistent with, but less resolved than trees inferred from combined molecular data. Parsimony mapping of morphological characters on the 3-gene trees showed that most structural features of steinernematids are highly homoplastic. The distribution of nematode foraging strategies on these trees predicts that S. hermaphroditum, S. diaprepesi and S. longicaudum (US isolate) have cruise forager behaviours.

Amino Acid Sequence↗

PubMed2006

Comparison of risk factors for candidemia versus bacteremia in hospitalized patients.

BACKGROUND: Classic risk factors for candidemia include use of total parenteral nutrition (TPN), hospital location, use of central venous catheter, and others. Unfortunately, most of these variables are now also risk factors for antibiotic-resistant bacteria. Thus, use of these risk factors to identify patients at high risk for candidemia is difficult. The purpose of this study was to compare these classic risk factors for candidemia in patients with bloodstream infections to determine the relative strength of these predictors in differentiating patients with candidemia and bacteremia. METHODS: Clinical data were collected from the medical charts of patients who had been hospitalized between 2002 and 2004. Patients with their first episode of candidemia or bacteremia during their hospital stays were included. Risk factors were assessed using a multivariate logistic regression model and internally validated using a bootstrap analysis. A p-value < 0.05 was considered significant. RESULTS: A total of 164 patients (82 with candidemia) were evaluated. According to the logistic analysis, patients who had stayed in the intensive care unit (ICU) (OR = 6.24; 95% CI: 2.58-15.09) or had been using TPN (OR = 4.69; 95% CI: 1.76-12.48) were more likely to have candidemia than bacteremia. While patients with pulmonary (OR = 0.15; 95% CI: 0.055-0.39) or cardiac disease (OR = 0.21; 95% CI: 0.086-0.51) had a greater chance to have bacteremia than candidemia (p < 0.01 for all variables). These results were further validated using bootstrap analysis. CONCLUSION: Among classic risk factors for candidemia, the ICU location at the time of culture and TPN use were most predictive of candidemia while certain medical disorders predicted patients at the highest risk for bacteremia. These results can be used to help identify patients most likely to benefit from empiric antifungal therapy.

Adult↗

PubMed1994

When prosody fails to cue syntactic structure: 9-month-olds' sensitivity to phonological versus syntactic phrases.

According to prosodic bootstrapping accounts of syntax acquisition, language learners use the correlation between syntactic boundaries and prosodic changes (e.g., pausing, vowel lengthening, large increases or decreases in fundamental frequency) to cue the presence and arrangement of syntactic constituents. However, recent linguistic accounts suggest that prosody does not directly reflect syntactic structure but rather is governed by independent prosodic units such as phonological phrases. To examine the implications of this view for the prosodic bootstrapping hypothesis, infants in Experiment 1 were presented with sentences in which pauses were inserted either between the subject noun phrase (NP) and verb or after the verb. Half of the infants heard sentences with lexical NP subjects, in which prosodic structure is consistent with syntactic structure. The other half heard sentences with pronoun subjects, in which prosodic structure does not mirror syntactic structure. In a preferential listening paradigm, infants in the lexical NP condition listened longer to materials containing pauses between the subject and verb, the main syntactic constituents. However, in the pronoun NP condition, infants showed no difference in listening times for the two pause locations. To determine if other sentence types containing pronoun subjects potentially provide information about the syntactic constituency of these elements, infants in Experiment 2 heard yes-no questions with pronoun subjects, in which the prosodic structure reflects the constituency of the subject. Infants listened longer when pauses were inserted between the subject and verb than after the verb. Taken together, our results suggest that the prosodic information in an individual sentence is not always sufficient to assign a syntactic structure. Rather, learners must engage in active inferential processes, using cross-sentence comparisons and other types of information to arrive at the correct syntactic representation.

Arousal↗

PubMed1983

How much evolutionary advantage does sex confer?

In discussing the long term advantage of sex, Crow and Kimura (1965) and Maynard Smith (1971) have argued that the advantage of a reproductive system allowing recombination (sex) is greatest for large populations. However the validity of this conclusion depends upon the model used for evolution. We propose two simple models: the bootstrap model, where the number of loci at which favourable mutations may take place remains constant over long time periods; and the environment-led model, where evolution is at a constant rate dictated by the environment (and does not depend on the organism's ability to evolve). While the bootstrap model leads to conclusions similar to those mentioned above, the conclusions for environment-led evolution are the opposite: as the size of the population decreases the advantage for sex increases.

Biological Evolution↗

PubMed1993

Higher-order structure in regularity detection.

In three experiments a simple Euclidean transformation (reflection, translation, rotation) was applied to collections of twelve dots in such a way that they contained equal lower-order structure, defined on the pairwise grouping of elements with their partner following transformation (e.g. parallel virtual lines), but differed in the presence vs absence of higher-order structure, defined on pairs of pairwise groupings (e.g. virtual quadrangles with correlated angles). Based on the much better performance levels (d') in the case of additional higher-order structure, we conclude that global regularities are easier to detect when the local correspondences are supported by higher-order ones formed between them. These enable the lower-order groupings to spread out across the whole pattern very rapidly (called bootstrapping). As a preliminary attempt to specify these principles, we proposed a working model with two basic components: first, a function expressing the cost of a perceptual grouping or the lack of regularity, and, secondly, an algorithm based on simulated annealing to minimize the cost function. The simulation results obtained with our current implementation of these principles showed satisfactory qualitative agreement with human regularity detection performance. Finally, the theory was shown to capture the essence of a large number of grouping phenomena taken from diverse domains such as detection of symmetry in dot patterns, global structure in Glass and vector patterns, correspondence in stereoscopic transparency and apparent motion. Therefore, we are convinced that, in principle, the mechanism used by the human visual system to detect regularity incorporates something like bootstrapping based on higher-order structure. We regard this as a promising step towards unraveling the intriguing mechanisms of classic Gestalt phenomena.

Cues↗

PubMed1992

Oddball-evoked P300-based method of deception detection in the laboratory. II: Utilization of non-selective activation of relevant knowledge.

We used a deception paradigm modeled on the type used for pre-employment screening procedures. Our novel dependent measure was P300 amplitude. Event-related potentials (ERPs) were recorded while subjects were presented with a list of eight antisocial acts one at a time, and one target-response phrase to which a 'yes' button press was required. Subjects were instructed to try to escape detection during the ERP test if they were guilty of any of the acts. After the ERP test, ground truth was established by the completion of an innocent/guilty check list of antisocial acts under perceived anonymous conditions tending to favor honest responding. Subjects were classified as innocent (n = 14) or guilty (n = 17) based on their check list response to the relevant act 'Used Falsified ID'. When comparing the P300 amplitudes in response to the relevant and to another act, we found that most group analyses revealed significant differences between guilty and innocent subjects. The subjects were also individually classified by a 3-step algorithm which involved: (1) a bootstrap amplitude test that compared the bootstrapped amplitudes of the P300s to the relevant and to another act; (2) relevant-to-target item P300 amplitude ratios; and (3) relevant act P300 amplitudes. Overall, the algorithm yielded 87% accuracy. The present study was intended to be an advance over our previous study (Rosenfeld, et al., 1991), in which we correctly classified 89% of the subjects using a similar P300-based deception detection paradigm. However, the possible confounding limitation of that study was that subjects had to complete an innocent/guilty check list of their antisocial acts prior to the ERP test. The present study investigated the accuracy of the P300-based test when subjects did not admit or selectively rehearse their guilt of the relevant act prior to the ERP test.

Electrodes↗

PubMed1996

Near-infrared spectroscopic characterization of pharmaceutical powder blends.

Near-infrared (near-IR) spectroscopy was used to qualitatively assess the homogeneity of a typical direct compression pharmaceutical powder blend consisting of hydrochlorothiazide, fast-flo lactose, croscarmellose sodium, and magnesium stearate. Near-IR diffuse reflectance spectra were collected from thieved powder samples using a grating-based spectrometer. A second-derivative calculation and principal component analysis were performed on the spectra prior to qualitative evaluation. Blend homogeneity was determined using single- and multiple-sample bootstrap algorithms and traditional chi-square analysis. The results suggested that bootstrap techniques provided greater sensitivity for assessing blend homogeneity than chi-square calculations and that near-IR has great potential as an analytical tool in powder blend analysis.

Carboxymethylcellulose Sodium↗

PubMed1992

18S rDNA sequences and the holometabolous insects.

The Holometabola (insects with complete metamorphosis: beetles, wasps, flies, fleas, butterflies, lacewings, and others) is a monophyletic group that includes the majority of the world's animal species. Holometabolous orders are well defined by morphological characters, but relationships among orders are unclear. In a search for a region of DNA that will clarify the interordinal relationships we sequenced approximately 1080 nucleotides of the 5' end of the 18S ribosomal RNA gene from representatives of 14 families of insects in the orders Hymenoptera (sawflies and wasps), Neuroptera (lacewing and antlion), Siphonaptera (flea), and Mecoptera (scorpionfly). We aligned the sequences with the published sequences of insects from the orders Coleoptera (beetle) and Diptera (mosquito and Drosophila), and the outgroups aphid, shrimp, and spider. Unlike the other insects examined in this study, the neuropterans have A-T rich insertions or expansion regions: one in the antlion was approximately 260 bp long. The dipteran 18S rDNA evolved rapidly, with over 3 times as many substitutions among the aligned sequences, and 2-3 times more unalignable nucleotides than other Holometabola, in violation of an insect-wide molecular clock. When we excluded the long-branched taxa (Diptera, shrimp, and spider) from the analysis, the most parsimonious (minimum-length) trees placed the beetle basal to other holometabolous orders, and supported a morphologically monophyletic clade including the fleas+scorpionflies (96% bootstrap support). However, most interordinal relationships were not significantly supported when tested by maximum likelihood or bootstrapping and were sensitive to the taxa included in the analysis. The most parsimonious and maximum-likelihood trees both separated the Coleoptera and Neuroptera, but this separation was not statistically significant.(ABSTRACT TRUNCATED AT 250 WORDS)

Animals↗

PubMed1995

Use of resampling techniques to estimate the variance of parameters in pharmacological assays when experimental protocols preclude independent replication: an example using Schild regressions.

Estimates of variance in pharmacological assays are usually made by repeating the experiment with different tissues. Biological factors, such as the inability to wash a drug from tissue, may preclude the type of replication that is appropriate for the statistics of interest. For example, in Schild regressions, replication is usually done at each concentration of antagonist. In some test systems, replication of dose-response curves is not possible. For example, some persistent agonists cannot be removed from tissues after exposure, while in other systems, rapid desensitization severely alters tissue sensitivity to repeated challenge with agonist. In this paper, we demonstrate how a statistical resampling method, bootstrapping, can be used to derive estimates of the confidence intervals for pA2, pKB, and slope from Schild plots. This method utilizes the speed of the computer to estimate variance by repeatedly resampling the data. The advantage to this method is that it can be used for many different experimental designs. For a data set obtained from a Schild regression of atenolol antagonism of isoproterenol in the guinea pig left atrium, bootstrap estimates of confidence limits were calculated for cases where dose ratios were derived from the same tissue and randomly paired tissues. These estimates showed good agreement with estimates obtained using conventional analytical methods, thus suggesting that this method may be useful in practice.

Adrenergic beta-Agonists↗

PubMed2006

Relationships at school and stage-environment fit as resources for adolescent engagement and achievement.

Guided by Self-Determination and associated theories, we examined whether adolescent (N=324, Mage=15, 52% female) competence (academic engagement and achievement) were supported by relationships at school and school fit. Aspects of relationships and school fit that were measured included adolescents' perceptions of each context as promoting autonomy, relatedness and competence. Within a latent-variable structural equation model, direct and indirect path estimates, standard errors and confidence intervals were produced using maximum likelihood and bootstrapping. Results supported the hypothesized model. As predicted, school fit partially mediated the association between teacher-student relationships and engagement, but fully mediated the association between peer relationships and engagement. Engagement fully mediated the path from school fit to achievement. The use of SEM and bootstrapping are encouraged as the combination of these techniques can increase power to detect direct and indirect effects, and can be a better choice for data that do not conform to normal theory assumptions. Overall, these techniques allowed for more firm conclusions about the importance of a hierarchy of multidimensional contextual experiences for adolescent competence.

Adaptation, Psychological↗

PubMed2004

A method for imputing missing data in longitudinal studies.

PURPOSE: In a cohort in which racial data are unknown for some persons, race-specific persons and person-years are imputed using a model-based iterative allocation algorithm (IAA). METHODS: An EM algorithm-based approach to address misclassification in a censored data regression setting can be adapted to estimate the probability that a person of unknown race is white. The corresponding race-specific person-years are obtained as a by-product of the estimation procedure. Variance estimates are computed using the bootstrap. The proposed approach is compared with the proportional allocation method (PAM). RESULTS: In an occupational cohort where racial data were missing for 41% of the workers, the age-time-race-specific person-years were estimated within a relative variation of approximately 20%, using the IAA. The deaths were less reliably estimated. The standardized mortality ratios (SMRs) for all-cause mortality estimated using the IAA and the PAM were more similar for the non-white workers than for a smaller subgroup of white workers. CONCLUSIONS: The IAA provides a method to reliably estimate race-specific person-year denominators in cohort studies with missing racial data. This method is applicable to other incompletely observed non-time-dependent categorical covariates. Internal cohort rates or SMRs can be computed and modeled, with bootstrap confidence intervals that account for the uncertainty in the determination of race.

Adolescent↗

PubMed2006

Confounder selection in environmental epidemiology: assessment of health effects of prenatal mercury exposure.

PURPOSE: The purpose of the study is to compare different approaches to the identification of confounders needed for analyzing observational data. Whereas standard analysis usually is conducted as if the confounders were known a priori, selection uncertainty also must be taken into account. METHODS: Confounders were selected by using backward elimination (BE), change in estimate (CIE) method, Akaike information criterion, Bayesian information criterion (BIC), and an empirical approach using a priori information. A modified ridge regression estimator, which shrinks effects of confounders toward zero, also was considered. For each criterion, uncertainty in the estimated exposure effect was assessed by using bootstrap simulations for which confounders were selected in each sample. These methods were illustrated by using data for mercury neurotoxicity in Faroe Islands children. Point estimates and standard errors of mercury effects on confounder-sensitive neurobehavioral outcomes were calculated for each selection procedure. RESULTS: The full model and the empirical a priori model showed approximately the same precision, and these methods were (slightly) inferior to only modified ridge regression. Lower precisions were obtained by using BE with a low cutoff level, BIC, and CIE. CONCLUSIONS: Standard analysis ignores model selection uncertainty and is likely to yield overoptimistic inferences. Thus, the traditional BE procedure with p = 5% should be avoided. If data-dependent procedures are required for confounder identification, we recommend that inferences be based on bootstrap statistics to describe the selection process.

Animals↗

PubMed2004

Predictors of prolonged air leak after pulmonary lobectomy.

BACKGROUND: The objective of this study was to identify the predictors of prolonged air leak (air leak longer than 7 days) in patients submitted to pulmonary lobectomy for lung cancer. METHODS: A retrospective analysis on 588 patients operated on of pulmonary lobectomy from January 1995 through June 2003 was performed. Univariate and logistic regression analyses were performed to generate a model predicting the risk of prolonged air leak. Bootstrap resampling technique was used to validate the regression model. RESULTS: A prolonged leak was exhibited by 15.6% of patients. Logistic regression analysis demonstrated that significant independent predictors of prolonged air leak were a reduced predicted postoperative forced expiratory volume in 1 second (p < 0.0001), the presence of pleural adhesions (p = 0.003), and upper resections (p = 0.006). Bootstrap resampling analysis confirmed the reliability of these variables. A regression equation was generated for the prediction of the risk of prolonged air leak. CONCLUSIONS: We report that a low predicted postoperative forced expiratory volume in 1 second, the presence of pleural adhesions, and the upper lobectomy or bilobectomy increased the risk of air leak persisting for more than 7 days. A model was generated to calculate this risk and assist the surgeon in taking extra measures to prevent such complication (ie, optimizing bronchodilator treatment, pleural tent, sealants, buttressed staple lines, water seal, and chest tube drainage).

Aged↗

PubMed2005

Modelling approach in cell/material interactions studies.

Based on our experiments, we propose a statistical modeling approach of the in vitro interactions between biological objects and materials. The objective of this paper is to provide basic principles for developing more ambitious experiments comparing the simultaneous influence of more than one or two parameters on various observations, taking advantage of convenient statistical and mathematical techniques for the treatment of measured data. Analyzing some examples of our own experiments, the essential features needed for modeling cell/material interaction studies are presented. Firstly, we describe the initial process of designing appropriate experiments that allow for comprehensive modeling. In the second part, we illustrate the different applications of a specific statistical modeling technique, the bootstrap protocol, on either the amplification of data, the elimination of correlation existing between measured parameters or, out of a set of parameters, identification of the most relevant parameter for further statistical analysis. Finally, based on recent statistical analysis tools such as the bootstrap, we illustrate the relative influence of biological and physical parameters in phenomenological studies of cell/material interactions.

Biocompatible Materials↗

PubMed2004

Musculoskeletal parameters of muscles crossing the shoulder and elbow and the effect of sarcomere length sample size on estimation of optimal muscle length.

BACKGROUND: Knowledge of musculoskeletal parameters is essential to understanding and modeling a muscle's force generating capability. A study of musculoskeletal parameters was conducted in two parts: (I) Empirical measurement of upper extremity musculoskeletal parameters. (II) Computational bootstrap simulation to examine statistical power of detecting optimal muscle length as a function of sarcomere length sample size and effect size. METHODS: Parameters were determined with a cadaver model. Sarcomere lengths were measured for 120 samples per muscle using laser diffraction and the mean sarcomere length used to estimate optimal muscle length. A bootstrap computational simulation was conducted to estimate variance in mean sarcomere length as a function of sample size. Statistical power for detecting optimal muscle length as a function of sample size and effect size was then determined. FINDINGS: Parameters are reported in tabular format. Power is 80% at approximately 85, 50, 40 and 25 samples for effect sizes of 0.5, 0.75, 1.0 and 1.5 mm respectively. INTERPRETATION: Musculoskeletal parameters for predicting muscle forces can be adequately measured in a cadaver model. Measurement of 40-60 sarcomere lengths per muscle is sufficient to calculate mean sarcomere length for estimating optimal muscle length with power of 80% for an effect size of 0.75-1.0 mm.

Adult↗

PubMed2005

A hospital perspective on the cost-effectiveness of beta-blockade for prophylaxis of atrial fibrillation after cardiothoracic surgery.

BACKGROUND: Prophylactic beta-blockade is the recommended strategy for suppressing atrial fibrillation after cardiothoracic surgery (CTS). However, beta-blockade's impact on the hospital length of stay (LOS) and other economic end points has not been adequately assessed. OBJECTIVE: The present evaluation sought to determine whether beta-blocker use after CTS is a cost-effective strategy for the prevention of postoperative atrial fibrillation (POAF). METHODS: This was a piggyback cost-effectiveness analysis of a prospective cohort evaluation comprising 1660 patients undergoing CTS at an urban academic hospital from October 1999 to October 2003. Patients receiving beta-blocker prophylaxis were matched 1:1 with control patients not receiving prophylaxis based on age >70 years, valvular surgery, history of atrial fibrillation, male sex, and use of preoperative digoxin or beta-blockers. The incidence of POAF, total hospital costs, and LOS were compared in each group. Nonparametric bootstrapping analysis was performed to examine the study results as part of a quadrant analysis and to calculate CIs for the incremental cost-effectiveness ratio. LOS and total costs were also compared in patients with and without POAF, regardless of beta-blocker use. RESULTS: Use of prophylactic beta-blockade was associated with a 17.3 % reduction in the incidence of POAF (P = 0.02) and a 2.2-day reduction in LOS (P = 0.001) compared with nonuse. It also was associated with a 25.7% reduction in total hospital costs compared with nonuse (mean [SD], $30,978 [$33,108] vs $41,700 [$67,369], respectively; P < 0.001), possibly due to a 27.6% reduction in room and board costs ($11,144 [$15,398] vs $14,920 [$22,132]; P < 0.001). In the bootstrapping analysis, 99.0% of the time prophylactic beta-blockade fell into quadrant IV, which indicated superior effectiveness and lower total costs. Regardless of beta-blocker use, patients who developed POAF had a significantly longer LOS compared with those who did not develop POAF (14.7 [19.1] days vs 10.1 [11.1] days, respectively; P < 0.001) and higher total costs ($47,240 [$85,941] vs $32,516 [$34,644]; P < 0.001). CONCLUSIONS: At the institution studied, beta-blocker prophylaxis against POAF after CTS was associated with significantly reduced total costs compared with nonuse of beta-blocker prophylaxis. Patients who developed POAF had significantly increased LOS and total costs compared with those who did not develop POAE An adequately powered prospective, randomized, placebo-controlled trial is necessary to confirm the results of this evaluation.

Adrenergic beta-Antagonists↗

PubMed2005

A resampling approach for adjustment in prediction models for covariate measurement error.

Recent works on covariate measurement errors focus on the possible biases in model coefficient estimates. Usually, measurement error in a covariate tends to attenuate the coefficient estimate for the covariate, i.e., a bias toward the null occurs. Measurement error in another confounding or interacting variable typically results in incomplete adjustment for that variable. Hence, the coefficient for the covariate of interest may be biased either toward or away from the null. This paper presents a new method based on a resampling technique to deal with covariate measurement errors in the context of prediction modeling. Prediction accuracy is our primary parameter of interest. Prediction accuracy of a model is defined as the success rate of prediction when the model predicts new response. We call our method bootstrap regression calibration (BRC). We study logistic regression with interacting covariates as our prediction model. We measure the prediction accuracy of a model by receiver operating characteristic (ROC) method. Results from simulations show that bootstrap regression calibration offers consistent enhancement over the commonly used regression calibration (RC) method in terms of improving prediction accuracy of the model and reducing bias in the estimated coefficients.

Calibration↗