Auto Categorization of twitter Users

8
Copyright 2008 Adobe Systems Incorporated. All rights reserved. Intelligent Categorization of Twitter Friends based on their Tweets Friends: Persons whom you follow Vamshi Krishna Reddy V IIITB Avinash Reddy Palleti IIITB

description

This explains about the automatic creation of lists based on friends tweets.

Transcript of Auto Categorization of twitter Users

Page 1: Auto Categorization of twitter Users

Copyright 2008 Adobe Systems Incorporated. All rights reserved.

Intelligent Categorization ofTwitter Friends based on their

Tweets

Friends: Persons whom you follow

Vamshi Krishna Reddy V IIITBAvinash Reddy Palleti IIITB

Page 2: Auto Categorization of twitter Users

Copyright 2008 Adobe Systems Incorporated. All rights reserved.

Introduction

In Twitter, you may be following several people, where you need to classify or organize your friends .

Twitter Lists feature is a way to organize the people you’re following on Twitter.

Example Lists can be: Art, Technology, Politics etc.,

Terminology: FRIENDS: People whom you follow on Twitter

Page 3: Auto Categorization of twitter Users

Copyright 2008 Adobe Systems Incorporated. All rights reserved.

Concept of Categorization of friends

As we have seen, the lists can be used to organize your friends.

Normally, we may want our friends to be organized automatically.

So there can be an auto list generator which will categorize the friends list for you based on the kind of tweets they tweet.

PROBLEM STATEMENT

The Auto list generator should be able to check the tweets of your friends and should categorize by identifying the genre they belong to.

Page 4: Auto Categorization of twitter Users

Copyright 2008 Adobe Systems Incorporated. All rights reserved.

Proposed Solution

Page 5: Auto Categorization of twitter Users

Copyright 2008 Adobe Systems Incorporated. All rights reserved.

Get the user ids of all friends

Page 6: Auto Categorization of twitter Users

Copyright 2008 Adobe Systems Incorporated. All rights reserved.

Retrieve tweets of each friend based on his user id

Page 7: Auto Categorization of twitter Users

Copyright 2008 Adobe Systems Incorporated. All rights reserved.

Extract the words that matter for us(Entities)

Page 8: Auto Categorization of twitter Users

Copyright 2008 Adobe Systems Incorporated. All rights reserved.

Finding out the Category to which our friend belong to

§Consider 4 words among extracted ones, say W, X, Y, Z. and each occurred 10,3,1,2 times respectively.

§Now we’ll calculate each category’s score as follows.

§S(Technology)= 10*Wtech + 3*Xtech + 1*Ytech + 2*Ztech

§Where Xtech is the probability of X co-occurring with Technology

§Similarly S(News), S(Music), S(Science) etc., are calculated.

§This calculation of S() is by considering all the words extracted from all the tweets of a given twitter user.

§Among all of categories which has got highest score will be the list into which the user belongs to.