Browse Code ▾
BOOTPM is a MatLab script that estimates transition probabilities and their standard deviations both theoretically and using the bootstrap technique.
It creates N samples sampling -with replacement- M observations and then computes the relative N TPM's.
Then, it simply computes the standard deviation of each transition probability using the distribution across all the N TPM's.
It also compares the boostrapped results to results estimated thoeretically, to detect "strange" results. It allows the user to choose a subsample of the data as well as an alternative specification of the states.
Editing the script, the user can activate different sections (to plot or save A and S, to change this function into an interactive script, to test the script, to compute other statistics, to customize it...)
Project Samples
