A Joint Model of Text and Aspect Ratings for Sentiment Summarization

1
A Joint Model of Text and Aspect Ratings for
Sentiment Summarization

• Ivan Titov (University of Illinois)
• Ryan McDonald (Google Inc.)
• ACL 2008

2
Introduction

• An example of an aspect-based summary
• Q1 Aspect identification and mention extraction
  (coarse or fine?)
• Q2 sentiment classification

3
Introduction Extraction problem
4
Assumptions for their model

• Ratable aspects normally represent coherent
  topics which can be potentially discovered from
  co-occurrence information in the text.
• Most predictive features of an aspect rating are
  features derived from the text segments
  discussing the corresponding aspect.

5
Multi-Aspect Sentiment model (MAS)

• This model consists of two pars
• Multi-Grain Latent Dirichlet Allocation (Titov
  and McDonald, 2008) build topics
• A set of sentiment predictors force specific
  topics correlated with a particular aspect.

6
MG-LDA (1)

• An extension of LDA (Latent Dirichlet
  Allocation) build topics that globally classify
  terms into product instances. (Creative Labs Mp3
  players versus iPods, New York versus Paris
  Hotels)
• MG-LDA models global topics and local topics.
• The distribution of global topics is fixed for a
  document, while the distribution of local topics
  is allowed to vary across the document.

7
MG-LDA (2)

• Ratable aspects will be captured by local topics
  and global topics will capture properties of
  reviewed items.
• Example . . . public transport in London is
  straightforward, the tube station is about an 8
  minute walk . . . or you can get a bus for 1.50
• A mixture of topic London (London, tube, )
• The ratable aspect location (transport, walk,
  bus)
• Local topics are reused between very different
  types of items.

8
MG-LDA (3)

• A doc is represented as a set of sliding windows,
  each covering T adjacent sentences.
• Each window v in doc d has an associated
  distribution over local topics and a
  distribution defining preference for local topics
  versus global topics A word can be sampled
  using any window covering its sentence s, where
  the window is chosen according to a categorical
  distribution
• Windows overlap permits the model to exploit a
  larger co-occurrence domain.
• Symmetrical Dirichlet prior for

9
Dirichlet distribution Dir(a)

• Its probability density function returns the
  belief that the probabilities of K rival events
  are xi given that each event has been observed ai
  - 1 times.

• Several images of the probability density of the
  Dirichlet distribution when K3 for various
  parameter vectors a. Clockwise from top left
  a(6, 2, 2), (3, 7, 5), (6, 2, 6), (2, 3, 4).

10
Multi-Aspect Sentiment Model (1)

• Assumption the text of the review discussing an
  aspect is predictive of its rating.
• MAS introduces a classifier for each aspect,
  which is used to predict its rating.
• Only words assigned to that topic can participate
  in the prediction of the sentiment rating of the
  aspect.
• However, rating for different aspects can be
  correlated. Ex. Negative cleanliness -gt rooms,
  service, dining.

11
Multi-Aspect Sentiment Model (2)

• Opinions about an item in general without
  referring to any particular aspect. Ex. This
  product is the worst I have ever purchased -gt low
  ratings for every aspect.
• Based on overall sentiment rating and compute
  corrections.
• N-gram model

12
Inference in MAS

• Gibbs sampling
• Appears only if ratings are known

13
Experiments - Corpus

• Reviews of hotels from TripAdvisor.com.
• 10,000 reviews (109,024 sentences, 2,145,313
  words in total)
• Every review was rated with at least 3 aspects
  service, location, and rooms.
• Ratings from 1 to 5.

14
Result Example
15
Evaluation

• 779 random sentences labeled with one or more
  aspects.
• 164, 176, 263 sentences for service, location,
  and rooms, respectively.

16
Results Aspect Service
17
Results Aspect Location
18
Result Aspect Rooms