Due to the large number of predictors available for analysis, only those interactions that were both biologically sensible and did not substantially inflate the standard errors of the estimates (i.e. had low multicollinearity), were considered for inclusion in the final model.
Thus, we took original integrative results from all 14 predictors for following analysis.
The first ten principal components were used as predictors for this analysis.
In order to ensure that the imputation model is at least as complex as the analysis model, and that the assumption of MAR is plausible, it is necessary to include the outcome variable and all possible likely predictors for the analysis model, in the imputation model.
The unadjusted models yielded 9 predictors for further analysis (P <0.25) with statistically significantly greater risk suggested for low maternal age, low maternal education, maternal non-working status, high EPDS score, low family income, high parity and statistically non-significant results for ethnicity (maternal), maternal stress, and infant sex.
All continuous predictors were centered for analysis (IRMA and woman-specific characteristics) and contextual sexual relevance was effect coded (1 = high sexual relevance, −1 = low sexual relevance).
HR-QoL is a proxy for health status, and an outcome variable of epidemiological, clinical, and health systems research studies; it is also an independent predictor for the analysis of the use and cost of health services [ 6, 7].
These curves have several identifying characteristics, which can be extracted quickly by splitting the input into several windows, obtaining the difference of their averages, and using these as predictors for the Random Forest analysis.
Therefore, specimen status, KRAS, BRAF and MSI were used as predictors for the multivariate analysis whereas exon 20 PIK3CA mutation was excluded as the sample size was less than 15.
The predictors for stepwise linear regression analysis were based on correlation analysis and selected from the variables known to be associated with the dependent variable.
