A REFERENCE-BASED RECOMMENDATION SYSTEM FOR ACADEMIC PAPERS ON - - PowerPoint PPT Presentation

a reference based recommendation system for academic
SMART_READER_LITE
LIVE PREVIEW

A REFERENCE-BASED RECOMMENDATION SYSTEM FOR ACADEMIC PAPERS ON - - PowerPoint PPT Presentation

A REFERENCE-BASED RECOMMENDATION SYSTEM FOR ACADEMIC PAPERS ON ACEMAP By Jingqi Zhang Recommendation systems are a useful alternative to search algorithms since they help users discover items they might be interested in. Recommendation


slide-1
SLIDE 1

A REFERENCE-BASED RECOMMENDATION SYSTEM FOR ACADEMIC PAPERS ON ACEMAP

By Jingqi Zhang

slide-2
SLIDE 2

Recommendation systems are a useful alternative to search algorithms since they help users discover items they might be interested in.

slide-3
SLIDE 3

Recommendation System in Acemap

...

slide-4
SLIDE 4

Recommendation System based on Author’s Name Problems

  • 1. Authors having the same

name

  • 2. The author has papers in

different fields

slide-5
SLIDE 5

Recommendation System based on Reference

l Based on Neighborhood Algorithm and Graph Theory l Strong relativity: 127 million paper + 530 million reference l High feasibility: can reach the speed of 60-100 recommended paper per second

slide-6
SLIDE 6

In this figure, A, B, C are different papers. A makes direct reference to B1, B1 makes direct reference to C1, so A makes indirect (two-step) reference to C1. The papers which are referenced by A are called A’s neighbors. For another paper D, we can also have D’s neighbors. The similarity of A and D is measured by numbers of the same neighbors they share.

A

B1 B2 C1 C2 C3 C4

reference

Neighborhood set of paper A

Neighborhood Algorithm

slide-7
SLIDE 7

Pricing for uplink power control in cognitive radio networks Spot pricing of secondary spectrum access in wireless cellular networks

6

...

3

reference reference

2

Game theory for cognitive radio networks: An

  • verview

Auction-based dynamic spectrum trading market — Spectrum allocation and profit sharing Recommendation Index r

Paper Network Based on Reference

slide-8
SLIDE 8

Neighborhood Algorithm based on Reference

Calculating Speed Searching Breadth Retrieval Precision

slide-9
SLIDE 9

Author’s Name-based vs. Reference-based

·Recommendation System based on Author’s Name ·Recommendation System based on Reference

Effectiveness improved compared to the current Acemap recommendation system.

slide-10
SLIDE 10

Infocom 2018 Recommendation System

slide-11
SLIDE 11

Multi-dimensional Recommendation

  • Most related papers
  • Most cited papers
  • Latest papers
  • Papers belonging to the same conference
  • Surveys for a specific field
slide-12
SLIDE 12

Visualization

Make Academia Visible!

slide-13
SLIDE 13

Session Map

The sessions on that map cluster naturally by the force of citation.

slide-14
SLIDE 14

THANK YOU