HMMs or Hidden Markov Models. I’ve always encountered this during the time I’m doing my undergrad thesis but I confess I haven’t really understood it fully. Nature Biotechnology has an interesting take on this here:
Surprisingly, HMM is first developed for speech processing. Interestingly, it is now an important tool in bioinformatics. (for info on its origin, see here Intro to HMM 1986)
In simple terms, HMM is about identifying patterns or observing how often a specific pattern or element occurs in a certain phenomenon. It kinda remind me of Sherlock Holmes’ Adventure of the Dancing Men (from Sir Arthur Conan Doyle’s work) where he cracked the code of the dancing men graffiti by observing how often a certain image appeared and connecting it to the letter “e” of the English alphabet.