early detection of twitter trends milan stanojevic [email protected] university of belgrade...
TRANSCRIPT
![Page 1: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/1.jpg)
EARLY DETECTION OF TWITTER TRENDS
MILAN STANOJEVIC
UNIVERSITY OF BELGRADE
SCHOOL OF ELECTRICAL ENGINEERING
![Page 2: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/2.jpg)
CONTENTS
Introduction
Trending topics
Parametric model
Data-Driven approach
Experiment results
Conclusion
2/22
Milan Stanojevic
![Page 3: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/3.jpg)
INTRODUCTION
Events occur in large datasets
We need: detection
classification
prediction
Parametric models are popular but overly simplistic
Nonparametric approach is proposed for time series inference
Observed signal is compared to two sets of reference signals – positive and negative examples
Is there enough information for earlier prediction?
(spoiler alert: YES)
3/22
Milan Stanojevic
![Page 4: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/4.jpg)
TRENDING TOPICS
Twitter: a global communication network
Tweet: a short, public message
Topic: a phrase in a tweet
Trending topic (trend): a topic that becomes popular
4/22
Milan Stanojevic
![Page 5: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/5.jpg)
PARAMETRIC MODEL
Expect certain type of pattern usually constant + jumps
Fit parameter in data e.g. size of a jump
5/22
Milan Stanojevic
![Page 6: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/6.jpg)
DATA-DRIVEN APPROACH
All the information needed is in the data
Assumptions: tweets are written by people
people are simple:
in how they spread information
in how they connect to each other
there is a small number of distinct ways in which a topic becomes trending
6/22
Milan Stanojevic
![Page 7: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/7.jpg)
DATA-DRIVEN APPROACH
7/22
Milan Stanojevic
![Page 8: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/8.jpg)
DATA DRIVEN APPROACH
8/22
Milan Stanojevic
![Page 9: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/9.jpg)
CLASSIFICATION BY EXPERTS
9/22
Milan Stanojevic
![Page 10: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/10.jpg)
CLASSIFICATION BY EXPERTS
10/22
Milan Stanojevic
![Page 11: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/11.jpg)
CLASSIFICATION BY EXPERTS
11/22
Milan Stanojevic
![Page 12: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/12.jpg)
CLASSIFICATION BY EXPERTS
12/22
Milan Stanojevic
![Page 13: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/13.jpg)
CLASSIFICATION BY EXPERTS
13/22
Milan Stanojevic
![Page 14: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/14.jpg)
CLASSIFICATION BY EXPERTS
14/22
Milan Stanojevic
![Page 15: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/15.jpg)
CLASSIFICATION BY EXPERTS
15/22
Milan Stanojevic
![Page 16: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/16.jpg)
CLASSIFICATION BY EXPERTS
16/22
Milan Stanojevic
![Page 17: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/17.jpg)
CLASSIFICATION BY EXPERTS
17/22
Milan Stanojevic
![Page 18: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/18.jpg)
CLASSIFICATION BY EXPERTS
Properties
simple: computation of distances
scalable: computation is easily parallelized
nonparametric: model “parameters” scale along with the data
18/22
Milan Stanojevic
![Page 19: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/19.jpg)
EXPERIMENT
SETUP
Dataset: 500 trends
500 non-trends
Do trend detection of 50% holdout set of topics
Online signal classification
RESULTS
Early detection 79% rate of early detection, 1.43hrs average
Low rate of error 95% true positive rate, 4% false positive rate
19/22
Milan Stanojevic
![Page 20: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/20.jpg)
EXPERIMENT
FPR / TPR Tradeoff
Early / Late Tradeoff
20/22
Milan Stanojevic
![Page 21: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/21.jpg)
CONCLUSION
New approach to detecting Twitter trends
Generalized time series analysis method: Classification
Prediction
Anomaly detection
Possible applications: Movie ticket sales
Stock prices
etc.
21/22
Milan Stanojevic
![Page 22: EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC sm123317m@student.etf.rs UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING](https://reader036.vdocuments.us/reader036/viewer/2022062516/56649da85503460f94a954f3/html5/thumbnails/22.jpg)
BIBLIOGRAPHY
Trend or No Trend: A Novel Nonparametric Method for Classifying Time Series
Stanislav Nikolov
Master thesis
Massachusetts Institute of Technology (2011)
22/22
Milan [email protected]