The use of LLMs to handle autocorrelation in time series data is really clever. It democratizes acces to complex statistical methods that used to require specialized training. Alex's metformin experiment is a great exaple of personal science in action, even if the effect is below perceptual thresold. The fact that the result remains significant after correcting for autocorrelation adds credibility.
The use of LLMs to handle autocorrelation in time series data is really clever. It democratizes acces to complex statistical methods that used to require specialized training. Alex's metformin experiment is a great exaple of personal science in action, even if the effect is below perceptual thresold. The fact that the result remains significant after correcting for autocorrelation adds credibility.