Acemap Research Paper Recommender System Based On Citation - - PowerPoint PPT Presentation

▶
acemap research paper recommender system based on citation
SMART_READER_LITE
LIVE PREVIEW

Acemap Research Paper Recommender System Based On Citation - - PowerPoint PPT Presentation

Feb 27, 2024 7 likes •290 views

Acemap Research Paper Recommender System Based On Citation Recommender System Examples Why&What : Research

slide-1
SLIDE 1

Acemap Research Paper Recommender System Based On Citation

slide-2
SLIDE 2
slide-3
SLIDE 3

Recommender System Examples

slide-4
SLIDE 4

Why&What : Research paper recommender system

slide-5
SLIDE 5

Why&What : Research paper recommender system

slide-6
SLIDE 6
slide-7
SLIDE 7

Our Approach : Based on citation Ø Same field of research Ø Similar problem interested in

  • Build a paper network consisting of 127 million

papers and 530 million reference relationships.

  • Offline algorithm
slide-8
SLIDE 8

Algorithm: Common Neighborhood method

D"→$ = &'()**'+, -'./'' )0 1 2) 3 4$ = +'5.ℎ7)/ℎ)), 8'2 )0 9:9'/ 3 |4$| = <=*7'/ )0 9:9'/8 in AA8 +'5.ℎ7)/ℎ)), 8'2

D"→$ = |BC ⋂ BE |

|BC |

slide-9
SLIDE 9
  • cites
  • Neighborhood set of paper A
  • Idea: Recommend papers with greater breadth
slide-10
SLIDE 10

How : Common Neighborhood method Example

  • Recommendation

degree

slide-11
SLIDE 11

Idea: Multidimensional recommendation matrix

slide-12
SLIDE 12

Idea: Multidimensional recommendation matrix

slide-13
SLIDE 13
slide-14
SLIDE 14
  • Current Acemap Paper Recommender System Analysis
slide-15
SLIDE 15

Recommendation result comparison

Recommender system based on authors' names Recommender system based on citation

Results of our algorithm are more convincing and reliable

slide-16
SLIDE 16

Only a few recommendation results if the authors has published only a few papers

Current Acemap Paper Recommender System Analysis

slide-17
SLIDE 17

Our algorithm can recommend multiple papers no matter how many papers the authors have published

Recommendation result comparison

slide-18
SLIDE 18

Results in different areas

  • 1. wireless communication
slide-19
SLIDE 19
  • 2. wireless network

Results in different areas

slide-20
SLIDE 20
  • 3. Recommendation area

Results in different areas

slide-21
SLIDE 21

Infocom 2018

http://acemap.sjtu.edu.cn/infocom2018

slide-22
SLIDE 22

Infocom 2018 Paper Recommendation Page

http://acemap.sjtu.edu.cn/infocomPaper?PaperID=p2735

slide-23
SLIDE 23

Infocom 2018 Affiliation Map

slide-24
SLIDE 24

Infocom 2018 Session Map

slide-25
SLIDE 25

Infocom 2018 Session Map

slide-26
SLIDE 26

Infocom 2018 Session Map

slide-27
SLIDE 27

Contribution

p Jiasheng Zhou

  • Prepared total paper dataset
  • Designed and implemented common neighborhood method
  • Implemented the idea of recommending papers with greater breadth
  • Did comparison experiments to evaluate proposed method
  • Participated in designing Infocom index page
  • Drawed Infocom 2018 session map and affiliation map

p Xinzhu Cai

  • Prepared total paper dataset
  • Designed and implemented common neighborhood method
  • Implemented the idea of multidimensional recommendation matrix
  • Optimized Infocom 2018 dataset
  • Designed Infocom recommendation result display page
  • Participated in drawing Infocom 2018 affiliation map
slide-28
SLIDE 28