Using Social Network Mining for Speech Behavior Analysis of Couples Sitting on a Sofa: (A Semantic Comparison between Happy and Unhappy Relationships)

Abstract

This study is an extension of our another research entitled “Using Process Mining for Predicting Relationships of Couples Sitting on a Sofa”, whereas the 5 most frequent/possible sitting positions for Happy Couples were identified/discovered as well as the 2 most frequent/possible sitting positions for Unhappy Couples. The main focus and emphasis of the current work is on Speech Behavior Analysis of the both Happy and Unhappy Couples, for each of the above-discussed sitting positions, in a semantic approach. To do this, 8 semantic keywords (i.e., in order to convey/represent the emotional status of the verbal words and phrases exchanged between the couples) were initially defined, and two Process Mining (process discovery) techniques/algorithms were later applied on the (previously collected) Sofa Data as the following: (1) Social Network Miner algorithm (based on the Subcontracting metric) supported by the ProM 6 Package Manager. (2) Fuzzy Miner algorithm (via frequency-based metric) supported by the Disco Fluxicon. Accordingly, the results showed that the occurrence of the keywords “Happy”, “Excited”, “Satisfied” and “In Love” was more frequent/possible in the following sitting positions: “Cuddling in the middle”, “Cuddling in the corner, “Side-by-Side (touching without cuddling)”, “Corner cuddle with tucked leggs” and “Legs on lap”. Alternatively, the occurrence of the keywords “Irritated”, “Sad”, “Angry” and “Worried” was more frequent/possible in the following sitting positions: “Opposite sides of the sofa” and “Sat on different sofas”. This study provides groundwork for further and future studies.