Abstract
It is well established that reviews contain expressions of sentiment
towards the reviewed items, e.g., "I liked it." Here, we hypothesize
that in the case of experience goods and specifically films, the re-
views also contain a signal of the emotions evoked by watching the
movie, emotions that were experienced when watching it. That
is, online reviews for experience goods also reflect the reviewer’s
emotions while experiencing the item, in the form of social shar-
ing and can be reliably extracted from it. We postulate that the
aggregated extracted emotional experiences for a movie form an
emotional signature that reflects the emotions evoked by the movie.
To establish that, we systematically conduct a set of analyses, each
designed to offer evidence that supports our hypothesis. The abil-
ity to reliably, efficiently, and unobtrusively obtain the emotions
evoked by films or other experience goods has numerous practi-
cal applications for both consumers and producers. For example,
affective recommender systems can incorporate the film’s evoked
towards the reviewed items, e.g., "I liked it." Here, we hypothesize
that in the case of experience goods and specifically films, the re-
views also contain a signal of the emotions evoked by watching the
movie, emotions that were experienced when watching it. That
is, online reviews for experience goods also reflect the reviewer’s
emotions while experiencing the item, in the form of social shar-
ing and can be reliably extracted from it. We postulate that the
aggregated extracted emotional experiences for a movie form an
emotional signature that reflects the emotions evoked by the movie.
To establish that, we systematically conduct a set of analyses, each
designed to offer evidence that supports our hypothesis. The abil-
ity to reliably, efficiently, and unobtrusively obtain the emotions
evoked by films or other experience goods has numerous practi-
cal applications for both consumers and producers. For example,
affective recommender systems can incorporate the film’s evoked
| Original language | English |
|---|---|
| Title of host publication | The 9th KDD Workshop on Issues of Sentiment Discovery and Opinion Mining |
| Pages | 1-10 |
| Number of pages | 10 |
| State | Published - 2020 |