Module 8: Sentiment Analysis (Applying ML to Text)

Our first applied deep-dive into unstructured data: turning raw movie-review text into features a classifier can learn from, and scaling the pipeline to datasets that don’t fit in memory.

What You’ll Learn¶

• The bag-of-words model and n-grams

• Term frequency–inverse document frequency (tf-idf) weighting

• Text cleaning and tokenization; stemming and stop words

• Training a logistic regression sentiment classifier on 50k IMDb reviews

• Out-of-core learning with stochastic gradient descent

• Topic modeling with latent Dirichlet allocation (LDA)

Lecture Materials¶

• 📓 Chapter notebook: Applying Machine Learning to Sentiment Analysis

Prerequisites¶

• Modules 3, 4, and 6

Reference¶

Raschka & Mirjalili, Machine Learning with PyTorch and Scikit-Learn, Chapter 8.