Recognizing Contextual Polarity in PhraseLevel Sentiment Analysis - PowerPoint PPT Presentation

Title: Recognizing Contextual Polarity in PhraseLevel Sentiment Analysis


1
Recognizing Contextual Polarity in Phrase-Level
Sentiment Analysis

• Theresa Wilson
• Janyce Wiebe
• Paul Hoffmann
• University of Pittsburgh

2
Introduction

• Sentiment analysis
• task of identifying positive and negative
  opinions, emotions, and evaluations
• How detailed? Depends on the application.
• Flame detection, review classification ?
  document-level analysis
• Question answering, review mining ?
  sentence or phrase-level analysis

3
Question Answering Example
Q What is the international reaction to the
reelection of Robert Mugabe as President of
Zimbabwe?
African observers generally approved of his
victory while Western Governments denounced it.
4
Prior Polarity versus Contextual Polarity

• Most approaches use a lexicon of positive and
  negative words
• Prior polarity out of context, positive or
  negative
• beautiful ? positive
• horrid ? negative
• A word may appear in a phrase that expresses a
  different polarity in context
• Contextual polarity

Cheers to Timothy Whitfield for the wonderfully
horrid visuals.
5
Example

• Philip Clap, President of the National
  Environment Trust, sums up well the general
  thrust of the reaction of environmental
  movements there is no reason at all to believe
  that the polluters are suddenly going to become
  reasonable.

6
Example

• Philip Clap, President of the National
  Environment Trust, sums up well the general
  thrust of the reaction of environmental
  movements there is no reason at all to believe
  that the polluters are suddenly going to become
  reasonable.

7
Example

• Philip Clap, President of the National
  Environment Trust, sums up well the general
  thrust of the reaction of environmental
  movements there is no reason at all to believe
  that the polluters are suddenly going to become
  reasonable.

Contextual polarity
prior polarity
8
Goal of Our Research

• Automatically distinguish prior and contextual
  polarity

9
Approach

• Use machine learning and variety of features
• Achieve significant results for a large subset of
  sentiment expressions

10
Outline

• Introduction
• Manual Annotations
• Corpus
• Prior-Polarity Subjectivity Lexicon
• Experiments
• Previous Work
• Conclusions

11
Manual Annotations

• Need sentiment expressions with contextual
  polarity
• ? positive and negative expressions of
  emotions, evaluations, stances
• Had subjective expression annotations in
  MPQA Opinion Corpus http//nrrc.mitre.org/NRRC/pub
  lications.htm
• ? words/phrases expressing
  emotions, evaluations, stances,
  speculations, etc.
• sentiment expressions ? subjective expressions
• Decision annotate subjective expressions in MPQA
  Corpus with their contextual polarity

12
Annotation Scheme

• Mark polarity of subjective expressions as
  positive, negative, both, or neutral

positive
African observers generally approved of his
victory while Western governments denounced it.
negative
Besides, politicians refer to good and evil
both
Jerome says the hospital feels no different than
a hospital in the states.
neutral
13
Annotation Scheme

• Judge the contextual polarity of sentiment
  ultimately being conveyed
• They have not succeeded, and will never succeed,
  in breaking the will of this valiant people.

positive
14
Agreement Study

• 10 documents with 447 subjective expressions
• Kappa 0.72 (82)
• Remove uncertain cases ? at least one annotator
  marked uncertain (18)
• Kappa 0.84 (90)
• (But all data included in experiments)

15
Outline

• Introduction
• Manual Annotations
• Corpus
• Prior-Polarity Subjectivity Lexicon
• Experiments
• Previous Work
• Conclusions

16
Corpus

• 425 documents from MPQA Opinion Corpus
• 15,991 subjective expressions in 8,984 sentences
• Divided into two sets
• Development set
• 66 docs / 2,808 subjective expressions
• Experiment set
• 359 docs / 13,183 subjective expressions
• Divided into 10 folds for cross-validation

17
Outline

• Introduction
• Manual Annotations
• Corpus
• Prior-Polarity Subjectivity Lexicon
• Experiments
• Previous Work
• Conclusions

18
Prior-Polarity Subjectivity Lexicon

• Over 8,000 words from a variety of sources
• Both manually and automatically identified
• Positive/negative words from General Inquirer and
  Hatzivassiloglou and McKeown (1997)
• All words in lexicon tagged with
• Prior polarity positive, negative, both, neutral
• Reliability strongly subjective (strongsubj),
  weakly subjective (weaksubj)

19
Outline

• Introduction
• Manual Annotations
• Corpus
• Prior-Polarity Subjectivity Lexicon
• Experiments
• Previous Work
• Conclusions

20
Experiments

• Both Steps
• BoosTexter AdaBoost.HM 5000 rounds boosting
• 10-fold cross validation

• Give each instance its own label

21
Definition of Gold Standard

• Given an instance inst from the lexicon
• if inst not in a subjective expression
• goldclass(inst) neutral
• else if inst in at least one positive and one
  negative subjective expression
• goldclass(inst) both
• else if inst in a mixture of negative and
  neutral
• goldclass(inst) negative
• else if inst in a mixture of positive and
  neutral
• goldclass(inst) positive
• else goldclass(inst) contextual polarity of
  subjective expression

22
Features

• Many inspired by Polanya Zaenen (2004)
  Contextual Valence Shifters
• Example little threat
• little truth
• Others capture dependency relationships between
  words
• Example
• wonderfully horrid

pos
mod
23

• Word features
• Modification features
• Structure features
• Sentence features
• Document feature

24

• Word token
• terrifies
• Word part-of-speech
• VB
• Context
• that terrifies me
• Prior Polarity
• negative
• Reliability
• strongsubj

• Word features
• Modification features
• Structure features
• Sentence features
• Document feature

25

• Word features
• Modification features
• Structure features
• Sentence features
• Document feature

• Binary features
• Preceded by
• adjective
• adverb (other than not)
• intensifier
• Self intensifier
• Modifies
• strongsubj clue
• weaksubj clue
• Modified by
• strongsubj clue
• weaksubj clue

Dependency Parse Tree
26

• Word features
• Modification features
• Structure features
• Sentence features
• Document feature

• Binary features
• In subject
• The human rights report poses
• In copular
• I am confident
• In passive voice
• must be regarded

27

• Word features
• Modification features
• Structure features
• Sentence features
• Document feature

• Count of strongsubj clues in
• previous, current, next sentence
• Count of weaksubj clues in
• previous, current, next sentence
• Counts of various parts of speech

28

• Document topic (15)
• economics
• health
• Kyoto protocol
• presidential election in Zimbabwe

• Word features
• Modification features
• Structure features
• Sentence features
• Document feature

Example The disease can be contracted if a
person is bitten by a certain tick or if a person
comes into contact with the blood of a congo
fever sufferer.
29
Results 1a
30
Results 1b
31
Step 2 Polarity Classification
19,506
5,671

• Classes
• positive, negative, both, neutral

32

• Word token
• Word prior polarity
• Negated
• Negated subject
• Modifies polarity
• Modified by polarity
• Conjunction polarity
• General polarity shifter
• Negative polarity shifter
• Positive polarity shifter

33

• Word token
• Word prior polarity
• Negated
• Negated subject
• Modifies polarity
• Modified by polarity
• Conjunction polarity
• General polarity shifter
• Negative polarity shifter
• Positive polarity shifter

• Word token
• terrifies
• Word prior polarity
• negative

34

• Word token
• Word prior polarity
• Negated
• Negated subject
• Modifies polarity
• Modified by polarity
• Conjunction polarity
• General polarity shifter
• Negative polarity shifter
• Positive polarity shifter

• Binary features
• Negated
• For example
• not good
• does not look very good
• not only good but amazing
• Negated subject
• No politically prudent Israeli could support
  either of them.

35

• Word token
• Word prior polarity
• Negated
• Negated subject
• Modifies polarity
• Modified by polarity
• Conjunction polarity
• General polarity shifter
• Negative polarity shifter
• Positive polarity shifter

• Modifies polarity
• 5 values positive, negative, neutral, both, not
  mod
• substantial negative
• Modified by polarity
• 5 values positive, negative, neutral, both, not
  mod
• challenge positive

36

• Word token
• Word prior polarity
• Negated
• Negated subject
• Modifies polarity
• Modified by polarity
• Conjunction polarity
• General polarity shifter
• Negative polarity shifter
• Positive polarity shifter

• Conjunction polarity
• 5 values positive, negative, neutral, both, not
  mod
• good negative

37

• Word token
• Word prior polarity
• Negated
• Negated subject
• Modifies polarity
• Modified by polarity
• Conjunction polarity
• General polarity shifter
• Negative polarity shifter
• Positive polarity shifter

• General polarity shifter
• pose little threat
• contains little truth
• Negative polarity shifter
• lack of understanding
• Positive polarity shifter
• abate the damage

38
Results 2a
39
Results 2b
40

• Ablation experiments removing features
• Negated, negated subject
• Modifies polarity, modified by polarity
• Conjunction polarity
• General, negative, positive polarity shifters

41
Outline

• Introduction
• Manual Annotations
• Corpus
• Prior-Polarity Subjectivity Lexicon
• Experiments
• Previous Work
• Conclusions

42
Previous Work

• Learn prior polarity of words and phrases
• e.g., Hatzivassiloglou McKeown (1997), Turney
  (2002)
• Sentence-level sentiment analysis
• e.g., Yu Hatzivassiloglou (2003), Kim Hovy
  (2004)
• Phrase-level contextual polarity classification
• e.g., Yi et al. (2003)

43
At HLT/EMNLP 2005

• Popescu Etizioni Extracting Product Features
  and Opinions from Reviews
• Choi, Cardie, Riloff Patwardhan Identifying
  Sources of Opinions with Conditional Random
  Fields and Extraction Patterns
• Alm, Roth Sproat Emotions from Text Machine
  Learning for Text-based Emotion Prediction

44
Outline

• Introduction
• Manual Annotations
• Corpus
• Prior-Polarity Subjectivity Lexicon
• Experiments
• Previous Work
• Conclusions

45
Conclusions

• Presented a two-step approach to phrase-level
  sentiment analysis
• Determine if an expression is neutral or polar
• Determines contextual polarity of the ones that
  are polar
• Automatically identify the contextual polarity of
  a large subset of sentiment expression

46
Thank you
47
Acknowledgments

• This work was supported by
• Advanced Research and Development Activity (ARDA)
• National Science Foundation