Markov Clustering on Person-to-Person Similarity Graph: Attribution of Movies’ Box Office Results to Preferences of Viewer Communities
- Title:
- Markov Clustering on Person-to-Person Similarity Graph: Attribution of Movies’ Box Office Results to Preferences of Viewer Communities
- Author(s):
- Tkachenko, Yegor
- Thesis Advisor(s):
- Jedidi, Kamel
- Date:
- 2013
- Type:
- Theses
- Degree:
- M.S., Columbia University
- Department(s):
- Business
- Persistent URL:
- https://doi.org/10.7916/D87M06G5
- Abstract:
- Search for methods of deriving actionable marketing segmentation has a long history in the marketing literature. This work proposes the use of Markov clustering algorithm on person-to-person similarity graph, where similarity between individuals is based on their similarity in rating assignments. This allows the detection of taste-based communities of users. Simple regression analysis is subsequently applied to detect the dependencies of box office results of movies of various genres on the preferences of specific viewer communities. The resulting analysis permitted identification of communities that drive box office results of specific movie genres.
- Subject(s):
- Business
Statistics
- Item views
- 1551
- Metadata:
-
text | xml
- Suggested Citation:
- Yegor Tkachenko, 2013, Markov Clustering on Person-to-Person Similarity Graph: Attribution of Movies’ Box Office Results to Preferences of Viewer Communities, Columbia University Academic Commons, https://doi.org/10.7916/D87M06G5.
Home