Baby Steps in Short-Text Classification with Python
Explore the fundamentals of short-text classification using Python in this EuroPython 2017 conference talk. Dive into a three-part presentation covering the approach to text classification problems, implementation of a Naive Bayesian model, and deployment in production environments. Learn about useful information sources, technology stack decisions, and solutions to common difficulties encountered during model training. Gain insights into alternative model choices like random forest and SVM. Examine a detailed architecture solution using REST calls between Java Client and Flask Server, and discover other deployment possibilities. Conclude with suggestions for model improvements and brief examples of supervised (CNN) and unsupervised (LDA) learning algorithms for text classification. Familiarize yourself with technologies such as Flask, Green Unicorn, uWSGI, NLTK, Sci-Kit, Python 3, Java 8, Jersey, Docker, and Kubernetes throughout this informative 31-minute talk.