Suchergebnisse

Abstract
Manual coding of text data from open-ended questions into different categories is time consuming and expensive. Automated coding uses statistical/machine learning to train on a small subset of manually-coded text answers. Recently, pretraining a general language model on vast amounts of unrelated data and then adapting the model to the specific application has proven effective in natural language processing. Using two data sets, we empirically investigate whether BERT, the currently dominant pretrained language model, is more effective at automated coding of answers to open-ended questions than other non-pretrained statistical learning approaches. We found fine-tuning the pretrained BERT parameters is essential as otherwise BERT is not competitive. Second, we found fine-tuned BERT barely beats the non-pretrained statistical learning approaches in terms of classification accuracy when trained on 100 manually coded observations. However, BERT's relative advantage increases rapidly when more manually coded observations (e.g., 200–400) are available for training. We conclude that for automatically coding answers to open-ended questions BERT is preferable to non-pretrained models such as support vector machines and boosting.

Occupation coding, an important task in official statistics, refers to coding a respondent's text answer into one of many hundreds of occupation codes. To date, occupation coding is still at least partially conducted manually, at great expense. We propose three methods for automatic coding: combining separate models for the detailed occupation codes and for aggregate occupation codes, a hybrid method that combines a duplicate-based approach with a statistical learning algorithm, and a modified nearest neighbor approach. Using data from the German General Social Survey (ALLBUS), we show that the proposed methods improve on both the coding accuracy of the underlying statistical learning algorithm and the coding accuracy of duplicates where duplicates exist. Further, we find defining duplicates based on ngram variables (a concept from text mining) is preferable to one based on exact string matches.