An Exploration of Text Analysis Methods to Identify Social Deliberative Skills

We report on a series of evaluations of text processing and machine learning methods with the goal of building classifiers for social deliberative skills, i.e. the capacity to deal productively with heterogeneous goals, values, or perspectives. Our corpus includes online deliberative dialogue from four diverse domain contexts. We use the LIWC and CohMetrix linquistic analysis tools to (1) assess differentiating characterisitcs of these domains, and (2) generte feature sets for machine learning. We report on comparative analysis of various machine learning alrorythms, feature selection methods, and training methods in our attempts to build classifiers for (1) individual skills and (2) a composite total