Dynamics of Love

Several research have tried to explain love and its time dynamics using mathematical models and patterns. This study aims to investigate the presence of time patterns specific for some core components of love relationships using a set of self-reported measures collected through a web platform. We hypothesized the emergence of differential patterns between people describing an ongoing relationship and people describing a past-relationship. Participants (N=480) were asked to fill in a survey, with generic information about themselves and their relationship. Then, individuals’ perceptions about the temporal dynamics of seven basic relational components (intensity, romance, passion, common interests, commitment, friendship, patience) were assessed by asking participants to draw, using a mouse or a touch screen device, a plot describing the fluctuation of each component in time, from the beginning of their relationship to “Today”. For every participant, the values of each component in different time windows were extracted. For each of the time window, one sum value was computed for every possible combination of the seven elements. On this summed values, different Machine Learning Models (Linear SVG, SGD Classifier) were trained to categorize the patterns in the two groups (current vs past).When tested on performances the models’ obtained probability scores higher than chance, with better classifications for relationship that lasted longer than 2 years. Findings suggest that the temporal patterns of relationships components have similar dynamics between different people and may therefore have some underneath physiological basis at the level of the central nervous system since individuals’ perceptions about their relationships are influenced by the context, that is their relationships status.

Preliminary results Poster for Summer Research Internship 2016, Nanyang Technological University, Singapore