Bayesian Essay Test Scoring sYstem - BETSY
Overview
  Bayesian approach
  Other essay scoring
       approaches
  Two Bayesian Models
  References

Papers
  JTLA (2002) paper

Betsy
  Help file (screen shots)
  Download
  Version History
  BETA Version
  Datasets

Demos
  Betsy
 

Other Naive Bayes Software Resources
   Mitchell
  Words
  NL Software

Text
Classification
  My favorite paper
  Links

Stemming
  Porter Paper
  Resources

Entropy
  Shannon (1948)
  Formulas
  Links

Machine
Learning
  Paper archive
  Resource Directory
  EU MLnet

Measurement
Decision Theory
  Methods explained

Links to Other Essay
Scoring Projects
  LSA 
  e-rater
  PEG
  Intellimetric
  ContentAnalyst
   

Bayesian
BEssay
BETest
BETScoring
BET sYstem

BETSY is a windows-based program that classifies text based on trained material. It was designed for automated essay scoring and can be applied to any text classification task.

Features include

  • Multinomial or Bernoulli Naive Bayes models

  • Optional Porter stemming

  • Re-entrant training

  • Popular database format

  • Output on screen and into CSV files

  • Training file trimming

  • Infrequent term purge

  • Web interface

  • Diagnostic information

  • Up to 5 score categories

  • Free for non-commercial use.

 

  

Betsy was developed by Lawrence M. Rudner, now of the Graduate Management Admission Council, formerly with Department of Measurement, Statistics and Evaluation, University of Maryland, College Park, MD 20742 with funds from the the U.S. Department of Education. Some evaluations of Betsy are being made possible with funds from the Maryland State Department of Education.