Evans, James and Pedro Aceves. 2016. “Machine Translation: Mining Text for Social Theory”. Annual Review of Sociology 42:21-50. DOI: 10.1146/annurev-soc-081715-074206
Given the difficulty in getting interpretable models from some of the ML methods - particularly deep learning-, what have researchers done to make statements about the features or the texts themselves? If these are used exclusively to predict labeling, What types of hypotheses have been answered this way?
Given the difficulty in getting interpretable models from some of the ML methods - particularly deep learning-, what have researchers done to make statements about the features or the texts themselves? If these are used exclusively to predict labeling, What types of hypotheses have been answered this way?