Reputation: 635
Next-word prediction or phrase-prediction engines used in modern keyboards of mobiles and tablets, like swift key & XT9, which predict the next word the user is going to type based on some pre-defined or dynamic corpus, based on n-grams (maximum frequency of last typed 2-3 words plus the current word) based language models (Markov Model).
What I think is that these engines/algos are a part of AI/NLP. But what I am not sure about is what specific branch of AI/NLP they belong to. Is it machine learning ? Is it data science ? Is it big data ? Is it Computing Intelligence ? Is it decision-making ? Is it data-mining ? Or statistical pattern recognition/ predictive analytics/ Supervised learning/ Unsupervised learning ? Or all/many of these or something else ?
Upvotes: 0
Views: 669
Reputation: 53
Word prediction techniques are well-established methods in the field of AAC (Augmentative and Alternative Communication) that are frequently used as communication aids for people with disabilities
1) accelerate the writing; 2) reduce the effort needed to type; 3) suggest the correct word (no misspellings)
source: http://medialab.di.unipi.it/web/Language+Intelligence/Prediction.pdf>
Upvotes: 0
Reputation: 9290
They tend to be called language models, and I'd say it's a branch of natural language processing.
I'd say that it is machine learning, good models will tend to use big data, and it is a supervised learning problem, though with a much different flavor than typical textbook supervised learning problems.
Upvotes: 7