Creating and testing specialized dictionaries for text analysis

Roman Taraban, Jessica Pittman, Taleen Nalabandian, Winson Fu Zun Yang, William M. Marcy, Srivinasa Murthy Gunturu

Research output: Contribution to journalArticlepeer-review

2 Scopus citations

Abstract

Practitioners in many domains-e.g., clinical psychologists, college instructors, researchers-collect written responses from clients. A well-developed method that has been applied to texts from sources like these is the computer application Linguistic Inquiry and Word Count (LIWC). LIWC uses the words in texts as cues to a person's thought processes, emotional states, intentions, and motivations. In the present study, we adopt analytic principles from LIWC and develop and test an alternative method of text analysis using naïve Bayes methods. We further show how output from the naïve Bayes analysis can be used for mark up of student work in order to provide immediate, constructive feedback to students and instructors.

Original languageEnglish
Pages (from-to)65-75
Number of pages11
JournalEast European Journal of Psycholinguistics
Volume6
Issue number1
DOIs
StatePublished - Jun 30 2019

Keywords

  • LIWC
  • Machine learning
  • Naïve Bayes
  • Text analysis

Fingerprint

Dive into the research topics of 'Creating and testing specialized dictionaries for text analysis'. Together they form a unique fingerprint.

Cite this