In a paper by Wainwright and Jordan on page 62 it mentions that a log partition function is **always** convex. This is done by showing that the second derivative of the log partition function is the covariance matrix of the sufficient statistic vector $phi(x)$.

Question is, is the covariance function guaranteed to be positive semi definite? With respect to the von-mises fisher distribution $proptoexp(kappamu^Tx)$ I get the second derivative to be w.r.t. $kappa$:

$$mu^Tleft(E(xx^T)-E(x)E(x)^Tright)mu$$

but, if I take the fact that the log partition function is

$$y=-log I_{nu}(kappa)+nulogkappa+(nu+1)log(2pi)\

y'=-frac{I_{nu+1}(kappa)}{I_{nu}(kappa)}+nufrac{1}{kappa}\

y''=frac{I_{nu+1}^2(kappa)-I_{nu}^2(kappa)}{I_{nu}^2(kappa)}+frac{2nu+1}{kappa}frac{I_{nu+1}(kappa)}{I_{nu}(kappa)}

$$

where $nu=d/2-1$ and plot the second derivative (for d=2), I get,

which is negative implying its concave.

Two possible explanations: 1. Covariance matrix is not necessarily positive definite. 2. numerical errors in plotting log bessel function

**so which assumption is wrong?**

**Contents**hide

#### Best Answer

Your expression for y' is wrong. Use $2I_nu'=I_{nu+1}+I_{nu-1}.$

### Similar Posts:

- Solved – Proving Ridge Regression is strictly convex
- Solved – Proof that CRF models and logistic models are convex functions
- Solved – Proof that CRF models and logistic models are convex functions
- Solved – Proof that CRF models and logistic models are convex functions
- Solved – Proof that CRF models and logistic models are convex functions