search

grambank / grambank-analysed

3 stars 0 forks source link

Set seeds? #78

Closed SamPassmore closed 2 years ago

SamPassmore commented 2 years ago

Hi Hedvig,

You mentioned at one point in the meeting yesterday that there is a random element to the building of the datasets that we analyse.

I wondered whether we should add set.seeds to those scripts so that the results should always be the same? This would mean that if there is a change, its because we coded something wrong or code changed somewhere, rather than it just being a coincidence?

HedvigS commented 2 years ago

Good call, adding that now. I'll add it to requirements.

It primarily matters for dialect reduction in the GB dataset and when pruning the tree.

SimonGreenhill commented 2 years ago

No! I hate these. what if you want to rerun the analysis to check inter-run variation?

HedvigS commented 2 years ago

No! I hate these. what if you want to rerun the analysis to check inter-run variation?

If so then you can just comment out that one line in the scriptset_random_seed.R. In the meantime, it's handy while we're collaborators in such different places to know that we're looking at the same output.