Toronto Metropolitan University
Browse
Behzadfar_Hesaneh.pdf (1.26 MB)

Applying supervised learning algorithms on information derived from Social Network to enhance recommender systems

Download (1.26 MB)
thesis
posted on 2021-05-24, 13:21 authored by Hesaneh Behzadfar
The aim of this research is to show how social networks can be used for marketing purposes. This is implemented with the assistance of learning algorithms. The method proposed in this research is based on the analysis of “Support Vector Machines”, which facilitates analysis of all information gathered from the social websites. It differs from other methods currently being used by social networking websites, which do not take advantage of classification. By using public information from social networks, a dataset was formed. It comprised of a thousand users and seven features. The examined features were location, age, gender, occupation, relationship status, and average travel time/year. In this research, the dataset will be examined twice: first using a regular SVM; and next by using “Weighted Feature Support Vector Machines”. For the latter, to assign weights, a method called “Pairwise Comparison” will be used to rank the importance of features.

History

Language

eng

Degree

  • Bachelor of Engineering

Program

  • Computer Science

Granting Institution

Ryerson University

LAC Thesis Type

  • Thesis

Thesis Advisor

Abdolreza Abhari