2020년 컴퓨터종합학술대회
Current Result Document : 12 / 12
| 한글제목(Korean Title) |
이웃기반 협업필터링에서 사전 필터링 기법들의 정확성 비교 |
| 영문제목(English Title) |
Comparison of Accuracy among the Pre-filtering Methods for Neighbor Selection in Collaborative Filtering |
| 저자(Author) |
Sao-I Kuan
김종민
송하주
Jongmin Kim
Ha-Joo Song
|
| 원문수록처(Citation) |
VOL 47 NO. 01 PP. 0862 ~ 0864 (2020. 07) |
한글내용
(Korean Abstract) |
|
영문내용
(English Abstract) |
Collaborative Filtering calculation costs most of the time in calculating with sparse data and similarity of neighbors. Choosing the suitable neighborhood set can save time and increase the accuracy, since correlation between the target user and the neighbors could affect quality directly for the recommendation system. We compared the result by using 4 different pre-filtering approaches to choose a similar neighbor set for collaborative filtering: Threshold filtering, Top-N filtering, Percentage filtering, and Negative-Percentage filtering. |
| 키워드(Keyword) |
|
| 파일첨부 |
PDF 다운로드
|
