Text mining online data with scikit-learn
summary
Uses Python 3 and scikit-learn to turn online text into features and shows what actually moves metrics: a count vectorizer bag‑of‑words baseline plus a support vector machine yields an F1 ≈ 0.527 on one classification task, while a Twitter spam set (100 spam / 398 not) produces 78.6% accuracy but F1 = 0 until a pipeline with grid search CV and n‑gram features raises precision to ~80% and recall to ~59%. Character n‑grams pick authors; LIWC‑style features estimate political stance.
AI-written summary · about
chapters (6)
- 0:00Introduction
- 3:17Count Vectorizer
- 6:53Batman Word Cloud
- 7:52Spam
- 11:24Parameterization
- 17:26Authorship
related talks
01h 24mMachine Learning with Text in scikit learn
01h 17mMachine Learning with Scikit-Learn
01h 26mBagging to BERT- A tour of applied NLP
42mThe Machine Understands Text — Learning and Search at Micro and Macro Scale
34mApplying Machine Learning Techniques to Detect and Understand Fake News
01h 09mAutomatic Tagging of Short Texts with scikit-learn and NLTK
35mHow does text become data?
01h 38mNavigating the World of NLP and Text Analytics