A corpus can be stored as a nested array of documents containing words in the order in which they occur in the text.
Topic De-Assignment
The topic assignment is removed from one word at a time. The likelihood that the word is assigned to any given topic will then be calculated. Here, probability of assignment is represented as the area of a 2D graph.
Probabilistic Selection
Once the probability of assignment to every topic has been calculated, a new topic is randomly selected.
Topic Updating
The word is then assigned the newly-determined topic and re-integrated into the corpus. Appropriate topic assignment tracking variables are likewise updated, and the next word in the corpus is de-assigned.
