Development of computational models for emotional diary text analysis to support maternal care

With the aim of supporting maternal care we propose new computational models for analyzing selfreported emotional diary texts of pregnant women. Based on experimentally gathered affective rating measures outside clinical setting we developed new models to facilitate interpretation and communication of affective expressions between persons representing different affectivity rating systems. We gathered affective ratings for 195 emotional adjectives and 16 pregnancy-related nouns in Finnish (n=35) in respect to three dimensions of pleasure, arousal and dominance. Our affectivity ratings showed accordance with some previous research results whereas comparison with collections of Finnish emotion words and affective ratings indicated some lack of overlap and discrepancies. Based on current experimental results we suggest further experiments in a real clinical setting to gather affective ratings from patients of maternal care to identify distinctive patterns in affective expressions indicating health needs to be addressed.