Model Selection Sample Clauses

Model Selection. In order to compare our dynamic latent trait model with the benchmark model we use the deviance information criterion (DIC; according to Xxxxxxxxxxxxx et al., 2002). The DIC is a generalization of the Akaike information criterion (AIC) and the Bayesian information criterion (BIC) for hierarchical models. In contrast to the AIC and BIC, DIC allows to compare Bayesian hierarchical models where the effective number of parame- ters is not clearly defined. Similar to the other information criteria a trade-off between model fit and model complexity is evaluated. The DIC contains one penalty term for the effective number of parameters used measuring model complexity and one term equal to the deviance of the likelihood measuring model fit. A lower DIC value indicates a better model fit. According to Xxxxxxxxxxxxx et al. (2002), if the difference in DIC is greater than 10, then the model with the larger DIC value has considerably less support than the model with the lower DIC value. For our models, the lower DIC value of our dynamic latent trait model (DIC = 9485.77) indicates that this model dominates in the terms of model fit as well as model complexity the obvious benchmark model (DIC = 12319.82). Results for the dynamic latent trait model Rating errors. We begin our analysis of the estimation results with the rating errors. Our dynamic latent trait model captures estimates for the rating bias µj and the standard deviation σj of the rating error of the big three external rating agencies on the score scale. Table 3.8 shows the results for the estimated posterior distribution of the parameters for the three raters µj and σj, respectively. The posterior distributions of the parameters are characterized by the mean values (mean) and the standard deviations (SD) of the 18, 000 (4 × 4, 500) posterior draws. We infer from Table 3.8 that Fitch has the smallest absolute rating bias from µj σj mean SD mean SD Fitch 0.0155 0.0018 0.0752 0.0021 − Moody’s 0.0887 0.0024 0.1013 0.0029 S&P 0.0732 0.0017 0.0641 0.0017 Table 3.8: Estimated rating bias µj and standard deviations σj for the rat- ing errors (on the score scale) of the big three external rating agencies Fitch, Moody’s and Standard&Poor’s. The posterior distributions of the parame- ters are characterized by the mean values (mean) and the standard deviations (SD) of the 18, 000 (4 × 4, 500) posterior draws. the consensus on the score scale with respect to the posterior mean (0.0155). Moody’s clearly seems to be too opt...
Sample 1Sample 2
AutoNDA by SimpleDocs
Model Selection. No major multicollinearity problems were detected for this initial logistic regression model. Two-way interaction was considered for prior antibiotic use by race and AGE contact outside the household, however, a likelihood ratio test for the interaction terms showed no statistically significant interaction (X2df=1=0.018, p=0.89). Since the interaction terms were found not to be statistically significant they were eliminated from the model. Assessment of confounding using the 10% change in estimate approach revealed no meaningful confounding by race or contact with an AGE affected person outside the household as all model subsets had an adjusted odds ratio within 10% of the full model (gold standard) (Supplemental Table 1). Precision of the odds ratio estimates was also considered and there was no meaningful gain in precision comparing the full model (CI ratio= 3.2) to the model with only prior antibiotic use (CI ratio=3.0) (Supplemental Table 1). Since there was little loss in precision when controlling for these variables it was decided that the model containing race and contact with an AGE affected person outside the household would be considered as the final model for norovirus-associated AGE. The model was found to have good fit with a deviance statistic of 3.58 and a p-value of
Sample 1
Model Selection. Using backwards elimination with a significance level of 0.10, the linear regression model for keratometric astigmatism at 1 year of age did not give any significant variables among surgical factors. Using the change variables, two surgical factors were given in the stepwise regression model: individual number of sutures (p = 0.061) and incision location (p = 0.071). The model is then: Δ K-Ast = 11.971 – 0.818 (Incision Loc.) – 0.830 (# Sutures). A mixed model with subject-specific intercepts containing all potential surgical factors of interest gave no significant results. The mixed model was given as: Ŷij = αi + bij tij + 0.213 X1 – 0.014 X2 + 0.061 X3 + 0.20 X4 – 0.14 X5 + 0.010 tij, where: Ŷij = Keratometric astigmatism in patient i at visit j; αi = Subject specific random effect on the intercept for subject i; bij = Subject specific random effect on the slope; X1 = Incision Type: 1 for Scleral Tunnel, 0 for Clear Corneal; X2 = Incision Location; X3 = Extended Keratome: 1 for Yes, 0 for No; X4 = Number of Sutures; X5 = Suture Type: 1 for Running, 0 for Interrupted; tij = Age of patient i at visit j, given in months. The lowest p-value of these variables was number of sutures (p = 0.190). The age variable in this model was also non-significant (p = 0.574). In the context of the model, all surgical factors are considered fixed effects, while age at surgery is treated as both a fixed and a random (subject-specific) effect.
Sample 1
Model Selection. The Delft3D suite of models will be utilized to provide a modelling platform for hydrodynamic and water quality modelling. A Delft3D model (“WHCW model”) covering marine waters of at least 7 km from the Project boundary has been developed in a previous preliminary study (1) (referred as the Feasibility Study hereafter) for developing the proposed CMPs. The WHCW model was developed based on the Update Model developed under the Update on Cumulative Water Quality and Hydrological
Sample 1
Model Selection. Table 4. Unadjusted and adjusted odds ratios of various characteristics with antenatal care adequacy Reproductive Age Women (15-49 years old) Receiving At Least One Prenatal Care Visit North and South Kivu (MICS 2010) Logistic regression analysis Crude Adjusted* Characteristic Odds Ratio 95% C.I.† p-value‡ Odds Ratio 95% C.I.† p-value‡ Locality <0.0001 0.0018 Urban 2.64 (1.62-4.29) 2.26 (1.36-3.77) Rural (reference) 1 -- 1 -- CRUDE MODEL: logit (P(D=1 | LOCALITY)) = b0 + b1*LOCALITY ADJUSTED MODEL: logit (P(D=1 | LOCALITY, AGE, PARITY, INTENTION, EDUCATION, RELIGION, REGION, WEALTH, MARITAL STATUS)) = b0 + b1*LOCALITY + b2*AGE + b3*PARITY + b4*INTENTION + b5*EDUCATION + b6*RELIGION + b7*REGION + b8*WEALTH + b9*MARITAL Age 0.226 0.4845 15-20 years old 21-26 years old 1.28 (0.64 - 2.57) 0.78 (0.55 - 1.11) 1.6 (0.71-3.62) 1.01 (0.64-1.59) 27+ years old (reference) 1 -- 1 -- CRUDE MODEL: logit (P(D=1 | AGE)) = b0 + b1*AGE ADJUSTED MODEL: logit (P(D=1 | LOCALITY, AGE, PARITY, INTENTION, EDUCATION, RELIGION, REGION, WEALTH, MARITAL STATUS)) = b0 + b1*LOCALITY + b2*AGE + b3*PARITY + b4*INTENTION + b5*EDUCATION + b6*RELIGION + b7*REGION + b8*WEALTH + b9*MARITAL Marital Status 0.213 0.4577 Never married 1.59 (0.74 - 3.44) 1.07 (0.50-2.30) married Currently 4.12) married 1 -- 1 -- (reference) Formerly 1.83 (0.81 - 1.71 (0.68-4.30) CRUDE MODEL: logit (P(D=1 | MARITAL)) = b0 + b1*MARITAL ADJUSTED MODEL: logit (P(D=1 | LOCALITY, AGE, PARITY, INTENTION, EDUCATION, RELIGION, REGION, WEALTH, MARITAL STATUS)) = b0 + b1*LOCALITY + b2*AGE + b3*PARITY + b4*INTENTION + b5*EDUCATION + b6*RELIGION + b7*REGION + b8*WEALTH + b9*MARITAL Parity 0.622 0.311 1-3 live births (reference) 1 -- 1 -- 4+ births 1.10 (0.77-1.57) 1.29 (0.79-2.11) CRUDE MODEL: logit (P(D=1 | PARITY)) = b0 + b1*PARITY ADJUSTED MODEL: logit (P(D=1 | LOCALITY, AGE, PARITY, INTENTION, EDUCATION, RELIGION, REGION, WEALTH, MARITAL STATUS)) = b0 + b1*LOCALITY + b2*AGE + b3*PARITY + b4*INTENTION + b5*EDUCATION + b6*RELIGION + b7*REGION + b8*WEALTH + b9*MARITAL Intendedness of Pregnancy
Sample 1
Model Selection. A logistic mixed effects model was used to investigate potential relationships between a binary outcome variable, i.e. the presence of a hitchhiker species with a group of explanatory variables such as manta ray gender. The model contained a random intercept to account for the correlation arising from individual mantas being repeatedly observed. To compare the goodness-of-fit, a GLMM model without random effects was tested. To ensure sufficient credibility to reliably estimate the parameters, categories of variables with cell counts below five were combined or removed such as injury type and breaching behaviour. The category ‘fresh mating wound’ from the pregnancy status variable was not included, since it was not possible to determine pregnancy status. The full model included the explanatory variables: manta ray gender, maturity status, pregnancy status, behavioural activity and sub-region (location of sighting). The Akaike information criterion (AIC) was used for the model selection procedure to determine the most important variables to include in the model. A lower AIC between two candidate models implies an improved fit to the data. The model was run separately for each of the hitchhiker species, and all of the variable combinations were tested (S2 Appendix, Tables 1-5). Next, the parameters (explanatory variables) with the lowest AIC were interpreted on the log odds scale (exp(parameter)) to obtain odds ratio values. The significance of each parameter was determined by whether the 95% confidence interval (CI) crossed one (non-significant). A narrow CI indicated that the estimate was known more precisely, in comparison to a wider CI which had a greater uncertainty. The analysis was performed using RStudio version 1.3.1056 [35].
Sample 1
Model Selection. The Fifth-Generation NCAR / Penn State Mesoscale Model Version 3.7 (MM5; Xxxxx et al. 1994) and the NCAR Advanced Research Weather Research and Forecasting Model Version 3 (WRF; Xxxxxxxxx et al. 2008) were selected as the two meteorological models to be implemented in the upgraded PATH modelling system. Preprocessor programs of the MM5 modelling system including terrain, REGRID, LITTLE_R, and INTERPF were used to develop model inputs.
Sample 1
AutoNDA by SimpleDocs
Model Selection. A set of 13 models was fit to the data to examine the importance of year-specific apparent survival (S), reach transition probabilities (ψ, probability of a fish moving from Black Rocks to Westwater, and vice versa), and p’s (Table 1). The modeling strategy was a typical one where best estimates of p’s for increasingly complex models were estimated and followed by addition of other parameters (see Xxxxxxx et al. 2010 or more details). The top model in the set contained 45% of the AICc weight and had 70 estimable parameters including survival rates for each reach and year and as a function of TL and TL2, transition probabilities, and probabilities of capture for every year, reach, and state combination. The second-ranked model had 35% of total model weight and one fewer parameter (the TL2 term), with all else being the same. Because the signs of the survival terms in the top and second-ranked models were the same and those models contained the bulk of the total weight (80%), and presented essentially the same trends, only the top-ranked model was interpreted in this analysis. A model with year and reach specific survival rates (94 total parameters, model 11 in the set) received no weight and many survival parameters were not estimable. Humpback Chub Abundance Annual abundance estimates for adult Humpback Chub (>200 mm TL) were calculated for 1998- 2012 using the Xxxxxxx estimator in the robust design model in Program MARK. The annual abundance estimates for Humpback Chub ranged from 1,139 (2008) to 6,747 (1998; Figure 2). Point estimates 95% confidence intervals (CI) for 1998–2000 were: 6,747 (4,001–11,636), 3,520 (2,513–4,979), and 2,266 (1,742–2,975), respectively. Point estimates for 2003–2005 were: 2,520 (1,814–3,554), 2,724 (2,034–3,689), and 2,000 (1,596–2,530), respectively. Point estimates for 2007–2008 were: 1,212 (972–1,532) and 1,139 (954–1,379), respectively. Point estimates for 2011–2012 were: 1,467 (1,175–1,861) and 1,315 (1,022–1,713), respectively (Figure 2). Significance of differences in estimates was tested based on over lapping confidence intervals (Schenker and Gentleman 2001). The last four years (2007, 2008, 2011, and 2012) were significantly (p<0.05) lower than the previous six years sampled (1998, 1999, 2000, 2003, 2004, and 2005) except for 2000, 2003, and 2005. Abundance estimates for juvenile Humpback Chub and first year adult Humpback Chub (200– 220 mm TL) were not attempted due to the low numbers of these size classes collected...
Sample 1
Model Selection 
Sample 1

Related to Model Selection

  • Panel Selection 1. The Parties shall apply the following procedures in selecting a Panel: (a) the Panel shall comprise 3 members; (b) within 15 days following the date of the establishment of the Panel, each Party shall nominate a Panelist; (c) the Parties shall endeavor

  • Shift Selection In multiple shift operations, employees within each classification shall have a right to select their work shift on the basis of their seniority within a bureau or division thereof and competing only with employees covered under this agreement on the following basis:

  • Provider Selection To the extent applicable to Provider in performance of the Agreement, Provider shall comply with 42 CFR 438.214, as may be amended from time to time, which includes, but is not limited to the selection and retention of providers, credentialing and recredentialing requirements and nondiscrimination. If Subcontractor and/or Health Plan delegate credentialing to Provider, Subcontractor and/or Health Plan will provide monitoring and oversight and Provider shall ensure that all licensed medical professionals are credentialed in accordance with Health Plan’s and the State Contract’s credentialing requirements.

  • Single Source Selection Services for tasks in circumstances which meet the requirements of paragraph 3.10 of the Consultant Guidelines for Single Source Selection, may, with the Association's prior agreement, be procured in accordance with the provisions of paragraphs 3.9 through 3.13 of the Consultant Guidelines.

  • Least-cost Selection Services for assignments which the Association agrees meet the requirements of paragraph 3.6 of the Consultant Guidelines may be procured under contracts awarded on the basis of Least-cost Selection in accordance with the provisions of paragraphs 3.1 and 3.6 of the Consultant Guidelines.

  • Vacation Selection Beginning January 2 of each calendar year, employees will be scheduled a time, based on seniority, to select up to three (3) segments of available vacation leave during the time period of April 1 through March 31. A “segment” is one (1) or more contiguous days of vacation leave. No segment shall include more than ten (10) consecutive days of vacation leave in June, July, and/or August, provided that an employee may select contiguous segments of vacation leave. Each employee will be guaranteed one (1) scheduled workweek of vacation leave if requested as one of their segments. Off-shift times to select a vacation will not be considered as “time worked” for purposes of computing callback or overtime. If an employee is unable to be present during their scheduled time they may make their choice by telephone, email, or another individual with written documentation of designation, may select a vacation segment(s) for the employee. If the employee fails to select their vacation during their assigned time, the Employer may proceed with scheduling. The employee will be provided an opportunity to select their segment(s) at a later date when they are available. The Employer will publish the vacation schedule by March 1, after considering requests, as well as Agency program needs. Employees will complete a Leave Request Form no less than thirty (30) days prior to any approved vacation segment taken.

  • Textbook Selection Textbook committees composed of teachers, principals, and the assistant superintendent shall review textbooks and shall make recommendations to the administrative team for consideration. Final selection shall be left to the Board of Education or its authorized representatives.

  • Quality-based Selection Services for assignments which the Bank agrees meet the requirements set forth in paragraph 3.2 of the Consultant Guidelines may be procured under contracts awarded on the basis of Quality-based Selection in accordance with the provisions of paragraphs 3.1 through 3.4 of the Consultant Guidelines.

  • Personnel Selection Leave 35.6.1 Where an employee participates in a personnel selection process for a position in the Public Service, as defined in the Financial Administration Act, the Council shall grant leave of absence with pay for the period during which the employee's presence is required for purposes of the selection process, and for such further period as the Council considers reasonable for the employee to travel to and from the place where his presence is so required.

  • Time Off for Selection Procedures A regular, limited-term or probationary employee shall be entitled to necessary time off with pay to participate in tests of fitness, examinations and interviews required by the Chief Human Resources Officer during working hours for the purpose of determining eligibility for movement to another class in the County service or transfer from one agency/department to another.

Time is Money Join Law Insider Premium to draft better contracts faster.