# Stochastic Differential Equation: minimum sample size in order quasi-likelihood estimation is reliable

Let I have a SDE model and I'm interested in estimation of its parameters from a sample of datas.

I would like to calculate the quasi-likelihood and estimate of the parameters of the Stochastic Differential Equation by the maximum likelihood method.

My question: what's the minimum sample size needed in order to estimate the parameters in a reliable way?

$50$ observations? $100$ observations? Or maybe $10,000$?

-