Search

Browse

• Communities
• Authors
• Year
• Titles
• Subjects

My Space

• Login

Affective analysis and modeling of spoken dialogue transcripts

Palogiannidi Elisavet

Full record

---

Summary

At this thesis we propose affective models for the emotional analysis of lexical units invarious granularity levels. We propose and evaluate the use of an affective-semantic modelto expand the affective lexica of German, Greek, English, Spanish and Portuguese. Moti-vated by the assumption that semantic similarity implies affective similarity, we use wordlevel semantic similarity scores as semantic features to estimate their corresponding af-fective scores. Various context-based semantic similarity metrics are investigated usingcontextual features that include both words and character n-grams. The model producescontinuous affective ratings in three dimensions (valence, arousal and dominance) for allfive languages, achieving consistent performance. We achieve classification accuracy (va-lence polarity task) between 85% and 91% for all five languages. For morphologicallyrich languages the proposed use of character n-grams is shown to improve performance.Moreover, we created the first Greek affective lexicon, translating the words of the Englishaffective lexicon ANEW and assigning them to native speakers for affective annotation. Itcontains human ratings for the three continuous affective dimensions of valence, arousaland dominance for 1034 words.Motivated by recent advances in the area of Compositional Distributional SemanticModels (CDSMs), we propose a compositional approach for estimating continuous affectiveratings for adjective-noun (AN) and noun-noun (NN) pairs. The ratings are computedfor the three basic dimensions of continuous affective spaces, namely, valence, arousal anddominance. We propose that similarly to the semantic modification that underlies CDSMs,affective modification may occur within the framework of affective spaces, especially whenthe constituent words of the linguistic structures under investigation form modifier-headpairs (e.g., AN and NN). The affective content of the entire structure is determined fromthe interaction between the respective constituents, i.e., the affect conveyed by the headis altered by the modifier. In addition, we investigate the fusion of the proposed modelwith the semantic-affective model proposed in literature applied both at word- and phrase-level. The automatically computed affective ratings were evaluated against human ratingsin terms of correlation. The most accurate estimates are achieved via fusion and absoluteperformance improvement up to 5% and 4% is reported for NN and AN, respectively.We also investigate text based models for the affective analysis of sentences that aremainly based on affective features. We investigate various datasets including news head-lines, movie subtitles, Twitter status updates and spoken dialogue transcriptions and thebest (state-of-the-art) perfor

Available Files

Services