Notice
Recent Posts
Recent Comments
Link
일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
1 | 2 | 3 | ||||
4 | 5 | 6 | 7 | 8 | 9 | 10 |
11 | 12 | 13 | 14 | 15 | 16 | 17 |
18 | 19 | 20 | 21 | 22 | 23 | 24 |
25 | 26 | 27 | 28 | 29 | 30 | 31 |
Tags
- git push
- git
- commit
- 1330
- 파이썬
- regExr
- boostcamp #aI tech #주간리뷰
- 함수
- 정규식
- 수정사항업데이트
- GitHub 설치
- 두수비교하기
- local repository 생성
- amend
- remote repository 생성
- 15596
- Restore
- github
- 백준
- 1
- Reset
- git commit
- STAGE
- git config global
- python
- ADD
- Baekjoon
- Push
Archives
- Today
- Total
Très bien
Content based filtering 본문
컨텐츠 기반 필터링 (Content based filtering)
컨텐츠 기반 필터링은 아이템(영화, 음악, 상품 등) 에 대한 설명과 사용자의 선호 profile (이전 구매 내역, 평점 등)에 기반하여 새로운 item을 추천합니다. 예를 들어, 특정 음악을 사용자가 선택하여 들었다면, 관련 가수, 장르, 키워드 등 item의 세부 컨텐츠가 유사한 다른 음악들을 추천할 수 있습니다.
- 장점 : Sparse data로부터 발생하는 Cold-start problem 에 잘 대응할 수 있습니다. 또한, 이력이 존재하지 않는 새로운 아이템의 추천 시나리오에 효과적입니다.
- 단점 : filter bubble이 생기면 이전 경험했던 아이템과 매우 유사한 특성의 아이템들만 제시하는 한계가 있으며, 새로운 사용자에 대한 추천에 한계가 있습니다.
Contents based Methods에서 사용할 수 있는 정보는 주로 다음과 같은 Contents 항목이 있습니다.
- User-side
- Demograpic information : gender, age, occupation, etc.
- Social networks
- Reviews, comments
- Item-side
- movie : genre, director, actor
- product : product images, product details, review, etc.
Reference
- boostcamp AI Tech Lecture Notes (RecSys), by 이성원
- NVIDIA Developer Technical Blog, How to Build a Winning Recommendataion System, by Carol McDonald Gabriel Moreira https://developer.nvidia.com/blog/how-to-build-a-winning-recommendation-system-part-1/
'Study > RecSys' 카테고리의 다른 글
(Paper Review) Deep Neural Networks for YouTube Recommendations (0) | 2022.03.19 |
---|---|
Recommender System (추천시스템) (0) | 2022.03.19 |
(Paper Review) Neural Collaborative Filtering (0) | 2022.03.19 |
Collaborative Filtering (협업 필터링) (0) | 2022.03.19 |
Movie Recommendation - Netflix Prize (0) | 2022.02.08 |
Comments