Please use this identifier to cite or link to this item:
http://hdl.handle.net/10609/109546
Title: | Predictions of new graph relationships - The Movie DataBase dataset |
Author: | Dolz Fernández, José Luís |
Director: | Hernández-González, Jerónimo |
Tutor: | Casas-Roma, Jordi |
Abstract: | Thanks to the huge amount of data that is collected nowadays, models can be created to make all kinds of predictions. Graphs are a speci c type of model that can connect this data through relationships and predict new ones. A clear example is the suggestions of new people to connect with in social networks. In this project, the information contained in The Movie Database of almost 5000 films from 1916 to 2017 is used to make a graph model and to predict brand new relationships: which actors will work together, who will be the director of a new blockbuster, etc. These new predictions are created by using machine learning over the relationships. The results obtained with best prediction algorithm used show an accuracy of 60%. Hence, further work is needed to tweak features extraction out from the graph model to improve the precision of these relationship predictions. |
Keywords: | graph analysis relationship prediction movies |
Document type: | info:eu-repo/semantics/masterThesis |
Issue Date: | 7-Jan-2020 |
Publication license: | http://creativecommons.org/licenses/by-sa/3.0/es/ |
Appears in Collections: | Bachelor thesis, research projects, etc. |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
jdolzfTFM0120memory.pdf | TFM memory | 1,17 MB | Adobe PDF | View/Open |
Share:
This item is licensed under a Creative Commons License