Description Usage Arguments Details Value Author(s) References See Also Examples
View source: R/IdentAnalysis.R
This function computes three types of practical identifiability scores, the smoothed condition number (SCN), the practical identifiability score (PIS), and Stanhope's kappa statistic.
1 | PISAnalysis(Y,S,L)
|
Y |
An |
S |
See details. |
L |
See details. |
This function computes three types of identifiability scores, SCN,
PIS, and Stanhope's kappa. Y is the matrix of discrete observations,
likely with measurement error. S, L are two nxn
-dimensional
matrices that represents the estimated inner product matrices between the
solution curves, and between the derivatives of solution curves and
the solution curves. In other words
\hat{Σ}_{\mathbf{x}\mathbf{x}} = Y S Y', \quad \hat{Σ}_{D\mathbf{x}, \mathbf{x}} = Y L Y'.
These two matrices are provided by function twostage2
.
SCN |
The smoothed condition number (SCN). |
PIS |
The practical identifiability score (PIS). |
kappa |
Stanhope's kappa statistic. |
Xing Qiu
Stanhope, S., Rubin, J. E., & Swigon, D. (2014). Identifiability of linear and linear-in-parameters dynamical systems from a single trajectory. SIAM Journal on Applied Dynamical Systems, 13(4), 1792-1815.
X. Qiu, T. Xu, B. Soltanalizadeh, and H. Wu. (2020+) Identifiability Analysis of Linear Ordinary Differential Equation Systems with a Single Trajectory. Submitted.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 | ## load Example 3.1. In this example, yy1 are discrete and noisy
## observations of (A2, x0.A), and yy2 are observations of (A2,x0.B).
## The first case is practically identifiable but the second is not.
data("example3.1")
## To demonstrate that PIS is useful, we will use the functional
## two-stage method to estimate A2
myfit1 <- twostage2(yy1, tt)
S <- myfit1$S; L <- myfit1$L
myfit2 <- twostage2(yy2, tt)
## PISs of the first case are all relatively large
## (good practical identifiability)
PISs1 <- PISAnalysis(yy1, S, L); PISs1
## PISs of the second case are all relatively small
## (bad practical identifiability)
PISs2 <- PISAnalysis(yy2, S, L); PISs2
## A2 is the true system matrix
round(A2,2)
## The first case is a good estimate of A2
round(myfit1$Ahat,2); round(sum((myfit1$Ahat - A2)^2),2)
## The second case is a bad estimate of A2 due to
## identifiability issues
round(myfit2$Ahat,2); round(sum((myfit2$Ahat - A2)^2),2)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.