#海外讲座直播#An Introduction to Text Mining

An Introduction to Text Mining

Mercury free and no after taste that some products have. Canadian pharmacy propecia? For a man with ED, a harder erection can lead to a more satisfying sexual experience. When it comes to hardness, this can help.

Katie Fawan & Mitch Fraas

 

What is text mining?

  • Collection of methods that involve computer-assisted analysis
  • Allow scholars to make connections and arguments about very large bodies of texts
  • Text mining can suggest patterns in language use

Types of text mining

  • Statistical methods            Sample: word frequency, basic topic modeling, word proximity
  • Natural language processing              Sample: Sentiment Analysis, Entity extracting

What questions can we ask?

  • Examine concepts over time, including histories
  • Find connect and aggregate significant terms in a corpus
  • Study characteristics of discourse based on a rage of factors(genre, gender, race, nationality, age)
  • Explore how event effect discourse and vice versa

How to practice?

  • Get a corpus(a bunch of  texts) , e.g. Oxford university continuing service ,hathi trust, wget, Textwranglar, Regexcoach, Screen Scraping,
  • Prepare your text (clean the data)
  • Choose and use an analytical tool
  • Analyze your results

Two key concepts

  • Token-- a collection of characters that a computer needs as the unit of value. Most text mining takes individual was as tokens, but there are a range of tools to manipulate what can learn about these tokens
  • Stopword list-- a collection of tokens that you as the program to ignore

Four tools

  • Find: Control-F. Command-F
  • Voyant  http://voyant-tools.org
  • Juxta
  • Topic Modeling (MALLET)    http://code.google.com/p/topic-modeling-tool

Understanding

  • Attend to how the tool works, how does it make results?
  • Tinker. often with text-mining, scholars return to the same tools and texts multiple times, shifting the parameters in order to see clearer or more nuanced results
  • Return to the text & context
  • Alternating close & distant reading
  • Reading result in the historical, geographical, or genre contexts that they are familiar with to form comparative analysis

Visualization

  • Textexure.conm
  • D3--- http://d3js.org

 


Comments are closed.



无觅相关文章插件