The Implementation of Douglas Biber’s Multi-Dimensional Approach to Text Classification

Abstract

The goal for this research project is to implement the text classification algorithm described by Douglas Biber in Variations across speech and writing. By creating this program, I will have created an automated tool to be used in and simply future research of text classification, allowing for the study of larger text samples. Biber’s algorithm identifies seven factors on which to classify text based on the patterns of co-occurrence between linguistic features. The program will read files and compute the factor scores for each text file, enabling the factor scores to be computed quickly and consistently. Once the program is created and accurately generates the factor scores, the resulting factor scores will be able to be used to search for correlations between text genre and other factors, for example how different groups of people might respond differently to various genres of text. This program will be a tool that can be used for many future research projects relating to text genres.