Date post: | 07-May-2015 |
Category: |
Education |
Upload: | marina-santini |
View: | 1,489 times |
Download: | 1 times |
MARINA SANTINI
P R O G R A M : C O M P U TAT I O N A L L I N G U I S T I C S A N D L A N G U A G E T E C H N O L O G Y
D E P T O F L I N G U I S T I C S A N D P H I L O L O G Y
UPPSALA UNIVERSITY, SWEDEN
21 NOV 2013
Semantic Analysis in Language Technology
Lecture 3 - Semantic-Oriented Applications:Sentiment Analysis
Course Website: http://stp.lingfil.uu.se/~santinim/sais/sais_fall2013.htm
Lecture 3: Sentiment Analysis
2
Acknowledgements
Thanks to Bing Liu for the many slides I borrowed from his Tutorial on Sentiment Analysis and Opinion Mining. Big thanks to Dan Jurafsky for his slides from Coursera NLP course.
Lecture 3: Sentiment Analysis
3
Lecture 3: Sentiment Analysis
4
Why are sentiments important (opinions/emotions/affects/attitudes/etc)
Lecture 3: Sentiment Analysis
5
Lecture 3: Sentiment Analysis
6
Lecture 3: Sentiment Analysis
7
Text Categorization Problem
Different level of granularity: Document Sentence Summary
Lecture 3: Sentiment Analysis
8
Lecture 3: Sentiment Analysis
9
Opionion: Formalization: Quadruple (4 components)
Lecture 3: Sentiment Analysis
10
Whatch out!
Date: The date is important in practice because one often wants to know how opinions change with time and opinion trends.
Lecture 3: Sentiment Analysis
11
Lecture 3: Sentiment Analysis
12
Lecture 3: Sentiment Analysis
13
Opionion: Formalization: Quintuple (5 components)
Lecture 3: Sentiment Analysis
14
Lecture 3: Sentiment Analysis
15
Lecture 3: Sentiment Analysis
16
Lecture 3: Sentiment Analysis
17
In which way ”sentiment” belongs to semantics?
Semantics is the study of meaning: It focuses on the relation
between signifiers, like words, phrases, signs, and symbols, and what they stand for. Through a semantics, we want to understand human language.
Through SA we want to automatically identify the meaning of certain words, phrases, etc. and how they relate to affective states expressed in texts (long, short, oral, written, etc.)
Lecture 3: Sentiment Analysis
18
Subjectivity & Emotion
Lecture 3: Sentiment Analysis
19
Subjectivity
Lecture 3: Sentiment Analysis
20
Emotion
Lecture 3: Sentiment Analysis
21
Sentiment, Subjectivity, Emotion
Lecture 3: Sentiment Analysis
22
Affect and Affective words…
http://research.microsoft.com/en-us/projects/tweetaffect/
Lecture 3: Sentiment Analysis
23
Lecture 3: Sentiment Analysis
24
Lecture 3: Sentiment Analysis
25
Basically… Text Classification!
Topic-based classificationGenre identificationAuthorship attribution
(plagiarism, authorship/classification of anonymous texts)
Spam filtersAutomatic email classification
(folder assignment)Threat identificationEtc.
Lecture 3: Sentiment Analysis
26
Lecture 3: Sentiment Analysis
27
Opinion Mining in the real world…
Lecture 3: Sentiment Analysis
28
UnSupervised Learning
Lecture 3: Sentiment Analysis
29
Supervised Classification
See Dan’s video presentation!
Lecture 3: Sentiment Analysis
30
Lecture 3: Sentiment Analysis
31
Lecture 3: Sentiment Analysis
32
Lecture 3: Sentiment Analysis
33
Lecture 3: Sentiment Analysis
34
Lecture 3: Sentiment Analysis
35
Lecture 3: Sentiment Analysis
36
Lecture 3: Sentiment Analysis
37
Lecture 3: Sentiment Analysis
38
Lecture 3: Sentiment Analysis
39
Team Work: 20 min; Discussion 15 min
You are going to apply for funding . You are interested in Horizion 2020 funding scheme (the new European research and innovation funding framework)
You think it is a good idea to create a Mood Index App.
Plan with your team mates this new sentiment-based app. Present to the audience the following aspects:
1) Purpose: what is the main use of this new app? (ex, identification of self-distructive behavior, depressive states, sad/happy mood, freindly attitudes, etc.)
2) Target users: who is going to use this app? (young people, parents, etc)3) Scenario: describe a typical scenario/context where your app is going to be used with
fruitful results4) Computational aspects: Which sentiment classes is the app going to identify? In which
language? Which computational model is going to be based upon?5) The actors: what kind of experts do you need? (ex a computational linguist, a app
developer, a psychiatrist, a company taking care of marketing and commercialization, a social worker, school teacher etc.)
6) Societal Benefits: How can the commercialization of your app contribute to decrease unemployment in your country and/or in EU.
7) Any additional aspect you might find relevant.
Lecture 3: Sentiment Analysis
40
How to build your own Twitter Sentiment Analysis Tool
http://blog.datumbox.com/how-to-build-your-own-twitter-sentiment-analysis-tool/
Lecture 3: Sentiment Analysis
41
This is the end… Thanks for your attention !