Word Prediction with N-Grams Model Using Python

The “AI611µ Word Prediction with N-Grams Model using Python” micro-course is about how to perform word prediction with artificial intelligence techniques. More precisely, it presents how to use N-Grams model to predict words based on a corpus and builds a small example with Python.

I gave this micro-course in 2020, once, at the ECAM Brussels Engineering School (ECAM) as a part of the “I4110 – Artificial intelligence” course. The course is taught in French, but all the material is available in English and in French.

Documents

• General information about the micro-course
• Competency Based Assessment
• Grid of skills to acquire

Theory

• Session 1: Word Prediction Problem and N-Grams Model
• Session 2: N-Grams Model Training and Model Evaluation

Practice

• Quizz 1: N-Grams model
• Quizz 2: Bigram model training
• Quizz 3: N-Grams model training and evaluation
• Coding 1: Corpus statistics
• Coding 2: Training a bigram model
• Mission 1: N-Grams model applications
• Mission 2: Bigram model training with nltk
• Project 1: Simple word prediction application

Resources

This section gathers resources that have been used to create this micro-course. These latter can be used to learn more about N-Grams.

Reference books

• Daniel Jurafsky and James H. Martin (2008). Speech and Language Processing (Second Edition). Pearson. (ISBN: 978-0-135-04196-3)

Online resources

• Online NGram Analyzer to perform simple statistics on texts.
• Google Ngram Viewer with data collected from Google Books.
• Official website of the Natural Language Toolkit (NLTK) Python module.