Implementasi Metode Collaborative Filtering Menggunakan Cosine Similarity dan Jaccard Similarity pada Sistem

Waskito, Muhammad Rizal (2024) Implementasi Metode Collaborative Filtering Menggunakan Cosine Similarity dan Jaccard Similarity pada Sistem. Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 18081010095.-cover.pdf Download (1MB)
	Text (BAB I) 18081010095.-bab1.pdf Download (106kB)
	Text (BAB II) 18081010095.-bab2.pdf Restricted to Registered users only until 20 September 2026. Download (111kB) \| Request a copy
	Text (BAB III) 18081010095.-bab3.pdf Restricted to Registered users only until 20 September 2026. Download (6MB) \| Request a copy
	Text (BAB IV) 18081010095.-bab4.pdf Restricted to Registered users only until 20 September 2026. Download (4MB) \| Request a copy
	Text (BAB V) 18081010095.-bab5.pdf Download (142kB)
	Text (Daftar Pustaka) 18081010095.-daftarpustaka.pdf Download (104kB)

Abstract

The rapid development of e-commerce has changed the way consumers shop, creating a need for more personalized and relevant product recommendation systems. The challenges that arise are often related to data sparsity, where recommender systems have difficulty providing accurate recommendations, especially when user data is limited. This study implements the Collaborative Filtering method with the Cosine Similarity and Jaccard Similarity algorithms in the context of a Single Page Application (SPA). Cosine Similarity is used to measure the similarity between users based on the value given to the product, while Jaccard Similarity focuses more on the similarity of user interactions without considering the rating value. The results show that Cosine Similarity tends to provide a higher similarity score than Jaccard Similarity, especially in situations with incomplete data, with an average score difference of 26.14%. The use of SPA in this e-commerce system also increases the interactivity and responsiveness of the user experience. In addition, the developed system utilizes the Fear of Missing Out (FoMO) effect to increase the urgency and relevance of product purchases. The integration of this algorithm, together with the SPA approach, not only improves the accuracy of recommendations but also opens up opportunities for further development, such as the application of hybrid filtering, to optimize the performance of the recommendation system in e-commerce.

Item Type:

Thesis (Undergraduate)

Contributors: