I happily report the successful completion of my Ph.D. candidacy exam obligations. I had an extremely hectic professional and personal schedule in the months leading up to the exam, which made my preparation more than challenging. Yet, I've made it through--no doubt due in part to a good and supportive exam committee. My best thank-you to them and to the world is to move forward.
And so, I want to record my thoughts as I develop the theory and pragmatics around my desired application of Markov models to Old English poetry. To start, I will review selected chapters from Foundations of Statistical Natural Language Processing (1999) by Christopher D. Manning and Hinrich Schütze. This text was recommended to me as a great place to begin developing an appropriately deep understanding of Markov models. I believe that I need to be very, very sharp on everything about Markov processes if my dissertation is to be as credible and useful, as I think it can be.
I'll begin, then, at Chapter 9, "Markov Models," the first chapter in the textbook's "Grammar" section. Because these are notes, I'll primarily use bullet lists to capture my observations and thinking.
- Hidden Markov Models (HMMs) comprise several variants in technique for statistical modeling.
- HMMs are widely used and have been generally successful for use in speech recognition.
- HMMs operate at the "hidden" structure of words in sentences; they allow us to look at the order of categories of words.
- This means that HMMs will be important in part-of-speech tagging.