Parametric Shared Frailty Models: Estimate
Estimate
An optional panel to specify the settings to control the estimation of the shared frailty models and the optional feature selection process.
- Alternating Direction Method or Multipliers (ADMM)
-
- Fast
- Applies the fast alternating direction method of multipliers (ADMM). This is the default.
- Traditional
- Applies the traditional ADMM algorithm.
- Apply L-1 regularization
- Conducts the process to control feature selection. The Penalty Parameter field specifies the penalty parameter that controls the regularization process. It must be a single value greater than 0. The default setting is 0.001.
- Model Convergence Criteria
-
- Parameter Convergence
- Specifies the convergence criteria for the parameter. It must be a single numeric value belonging to [0, 1). The default setting is 0.000001. For Type, you can select either ABSOLUTE to apply the absolute convergence to the inner optimization or RELATIVE to apply the relative convergence to the inner optimization. The optional Value specifies a numeric threshold for the convergence type.
- Objective Function Convergence
- Specifies the convergence criteria for the objective function. It must be a single numeric value that belongs to [0, 1). The default setting is 0, which does not apply the convergence criteria. For Type, you can select either ABSOLUTE to apply the absolute convergence to the inner optimization or RELATIVE to apply the relative convergence to the inner optimization. The optional Value specifies a numeric threshold for the convergence type.
- Hessian Convergence
- Specifies the convergence criteria for the Hessian matrix. It must be a single numeric value that belongs to [0, 1). The default setting is 0, which does not apply the convergence criteria. For Type, you can select either ABSOLUTE to apply the absolute convergence to the inner optimization or RELATIVE to apply the relative convergence to the inner optimization. The optional Value specifies a numeric threshold for the convergence type.
- Residual Convergence Criteria
- An option to control the optimization process.
- Both primal and dual residual
- Applies both primal and dual residual convergence criterion. This setting is by default.
- Primal residual only
- Applies the primal residual convergence criterion.
- Dual residual only
- Applies the dual residual convergence criterion.
- Method
- An optional parameter to specify the estimation method.
- Auto
- Automatically chooses the method based on the sample data set. This method is selected by default. The Threshold number of predictors field specifies the threshold of the number of predictors, and must be a single integer greater than 1. The default value is 1000.
- Newton-Raphson
- Applies the Newton-Raphson’s method.
- L-BFGS
- Applies the limited-memory BFGS algorithm. The Update field specifies the number of the past updates that are maintained by the limited-memory BFGS algorithm, and must be a single integer greater than or equal to 1. The default value is 5.
- Iteration
-
- Maximum iterations
- Specifies the maximum number of iterations. It must be a single integer that belongs to [1, 300]. The default setting is 20.
- Maximum step-halving
- Specifies the maximum number of step-halving. It must be a single integer that belongs to [1, 200]. The default setting is 5.
- Maximum number of line searches
- Specifies the maximum number of line searches. It must be a single integer that belongs to [1, 300]. The default setting is 20.
- Absolute convergence for iteration process
- Specifies the absolute convergence for the outer iteration process. It must be a single numeric value that belongs to (0, 1). The default setting is 0.0001.
- Relative convergence for iteration process
- Specifies the relative convergence for the outer iteration process. It must be a single numeric value that belongs to (0, 1). The default setting is 0.01.