Intro to Natural Language Text Mining - Short Course
This class will cover machine learning applied to natural language text documents. Applications include: text-based search, predicting page reads based on text content, sentiment analysis and automated page and email construction based on user history. We will cover the use of statistical algorithms for accomplishing machine learning tasks on texts.
The course starts with some introduction to the subject matter, comparison of statistical techniques to semantic approaches, definition of problems in text mining, and simple text manipulations. We'll go through some basics of the R-language so you can follow the code examples presented. We'll cover various algorithms for dealing with standard text mining problems, such as indexing, automatic classification (e.g. spam filtering) topic modeling, classification etc.
I. Intro to text mining problems
II. R language background
III Structuring Text
IV Document-Term Matrix Processing
-Formation and Basic Manipulations of Document-Term Matrix
-Latent Semantic Indexing - Search
-Topic Modelling - Clustering and Classification
Prerequisites - Programming experience is required. We'll use R code examples to work through the material. You should have R installed and R Studio. Here are links for those.
There will be a short intro to R for those who haven't used it.
Class will meet in two sessions: Wed 6/26 and Tues 7/2 from 6pm to 9 pm. Registration covers both sessions. Pay by credit card
or paypal (mike at mbowles dot com), or pay by check or cash at the first session. There's a $50 discount if you sign up at least 5 days before the class starts.
The class will be webcast for those who want to view remotely. To receive the webcast instructions, you'll need to sign up on eventbrite 24 hours before the start of class.