Buot, Alain Jared N.; Dizon, Kyle-Jie B. 4 0
Art Smart: an enhanced content-based filtering algorithm using levenshtein distance for art recommendation system / 6 6 Buot, Alain Jared N.; Dizon, Kyle-Jie B. - - - ix, 63 pp. 28 cm. - - - - - . - . - 0 . - . - 0 .
Undergraduate Thesis: (Bachelor of Science in Information Technology), Pamantasan ng Lungsod ng Maynila, 2023.
5
ABSTRACT: This paper attempts to modify the Content-Based Filtering Algorithm (one of the known algorithms in Recommender Systems) using the Levenshtein Distance for Art Recommendation System. Recommended Systems are vital in today's age since there is so much information available on the internet, and these systems are the ones in charge of filtering these massive amounts of data to fit the users interests. This paper focuses on some of the drawbacks of the Algorithm, one of which is Overspecialization, when the Algorithm recommends items to the user that are specified on a single field or type only. Second, is the algorithm's lack of diversity in which the recommended items are only derived from the user's positively rated items which restricts a more diverse recommendation. And the last drawback is how the algorithm shows bias to the top-rated items which gives newer and less popular items a cold start. The goals that were achieved in this study is first, to provide a more diverse recommendation based on the user's profile, second is to provide two kinds of recommendations that are based on the user's overall likes and recent likes. And lastly, to give all items an equal chance to be recommended regardless of the rating. The Researchers gathered the data from data.world, which consists of different information about each artwork and its artist. The findings imply that the modified algorithm's use has improved compared to the original. Just like the original Content-Based Filtering, it suggests artworks based on their previous interests to users. However, it also recommends fresh and familiar artworks that may expand users interest.
5
2 = =
2
2 --0------
6 --0-- 2 --------
0 2 --
--20------
--------20--
--------20--
----2
/ 2
/ 2
/
/
Art Smart: an enhanced content-based filtering algorithm using levenshtein distance for art recommendation system / 6 6 Buot, Alain Jared N.; Dizon, Kyle-Jie B. - - - ix, 63 pp. 28 cm. - - - - - . - . - 0 . - . - 0 .
Undergraduate Thesis: (Bachelor of Science in Information Technology), Pamantasan ng Lungsod ng Maynila, 2023.
5
ABSTRACT: This paper attempts to modify the Content-Based Filtering Algorithm (one of the known algorithms in Recommender Systems) using the Levenshtein Distance for Art Recommendation System. Recommended Systems are vital in today's age since there is so much information available on the internet, and these systems are the ones in charge of filtering these massive amounts of data to fit the users interests. This paper focuses on some of the drawbacks of the Algorithm, one of which is Overspecialization, when the Algorithm recommends items to the user that are specified on a single field or type only. Second, is the algorithm's lack of diversity in which the recommended items are only derived from the user's positively rated items which restricts a more diverse recommendation. And the last drawback is how the algorithm shows bias to the top-rated items which gives newer and less popular items a cold start. The goals that were achieved in this study is first, to provide a more diverse recommendation based on the user's profile, second is to provide two kinds of recommendations that are based on the user's overall likes and recent likes. And lastly, to give all items an equal chance to be recommended regardless of the rating. The Researchers gathered the data from data.world, which consists of different information about each artwork and its artist. The findings imply that the modified algorithm's use has improved compared to the original. Just like the original Content-Based Filtering, it suggests artworks based on their previous interests to users. However, it also recommends fresh and familiar artworks that may expand users interest.
5
2 = =
2
2 --0------
6 --0-- 2 --------
0 2 --
--20------
--------20--
--------20--
----2
/ 2
/ 2
/
/