Asee peer logo

Recommendation Engine using Adamic Adar Measure

Download Paper |

Conference

2022 Spring ASEE Middle Atlantic Section Conference

Location

Newark, New Jersey

Publication Date

April 22, 2022

Start Date

April 22, 2022

End Date

April 23, 2022

Page Count

17

DOI

10.18260/1-2--40066

Permanent URL

https://peer.asee.org/40066

Download Count

385

Paper Authors

biography

Sourabh Dadapure Sacred Heart University

visit author page

I’m a Software engineer at a tech startup. I’m also an experienced full-stack developer and a data analyst.
Apart from academics, I’m a public speaker, investor, mentor, and explorer. I’m also an avid reader and a guitar enthusiast, my favorite being "Atomic Habits" by James Clear. Having built technological advancements throughout my life, I also have a strong motivation to keep expanding my knowledge and share that knowledge to empower others

visit author page

Download Paper |

Abstract

ABSTRACT

In recent years, recommendation engines have gained a lot of success on many online giant commerce and entertainment platforms.

Recommending similar products that users will like using the user's past behavior is a challenging problem especially because of the unpredictable nature of people’s likes and dislikes. It also involves a guess about the future based on something that the user has never seen which makes it that much harder to predict mainly because people’s tastes change all the time. What we can do is try to estimate those values as best as we can using the Adamic Adar measure by creating nodes and finding similarities between those nodes. Unlike most of the existing recommendation systems that use either collaborative filtering or content-based filtering to generate recommendations, this paper explores a slightly different approach by creating node pairs consisting of common neighbors but with a lower degree and calculating the Adamic Adar Coefficient of those two nodes. Adamic Adar Coefficient is a measure that is used to calculate the closeness of two nodes based on their common neighbor. This paper describes a recommendation engine built that can be used to predict similar items when a user is browsing an eCommerce, music or movie platform based on the user’s behavior. It takes in the item’s features such as description, price, title, ratings, etc., and creates nodes for each word to find commonalities between those nodes. It then generates nodes with the highest Adamic Adar Coefficient which will result in the items that are close in characteristics to the currently viewed item by the user.

Adamic Adar Coefficient: If I and J are two nodes, the Adamic Adar Coefficient of I and J would be calculated as Whereas N(node) is a function that returns the set of neighboring nodes

Dadapure, S. (2022, April), Recommendation Engine using Adamic Adar Measure Paper presented at 2022 Spring ASEE Middle Atlantic Section Conference, Newark, New Jersey. 10.18260/1-2--40066

ASEE holds the copyright on this document. It may be read by the public free of charge. Authors may archive their work on personal websites or in institutional repositories with the following citation: © 2022 American Society for Engineering Education. Other scholars may excerpt or quote from these materials with the same citation. When excerpting or quoting from Conference Proceedings, authors should, in addition to noting the ASEE copyright, list all the original authors and their institutions and name the host city of the conference. - Last updated April 1, 2015