Transcript Japanese Weblog Opinion Mining

Opinion mining in social networks
Student: Aleksandar Ponjavić 3244/2014
Mentor: Profesor dr Veljko Milutinović
2/16
Introduction
• Opinion mining is a type
of natural language processing
for tracking the mood of the public
• Opinion mining involves building a system
to collect and categorize opinions
• Data – products, topic
3/16
Social networks
• Social networks are best represented as graphs
• Social power (member’s prestige) is centrality
• Centrality
▫ number of links
▫ number of shortest paths
▫ the mean of shortest paths lengths
4/16
Opinion mining
• The first task is sentiment analysis
and aims at the establishment of the polarity
of the given source text
• Some words have different meanings
in various contexts
5/16
Opinion mining
• The second task consists
in identifying the degree of objectivity
and subjectivity of a text
• Opinion extraction
6/16
Opinion mining
• The third task is aims at the discovery
and/or summarization of explicit opinions
of the assessed product.
• All three classes of opinion mining tasks
can greatly benefit from additional data
from the social network (centrality).
7/16
Document semantic orientation
•
•
•
•
Ti – the i-th term of the document d
|d| – is the number of terms appearing in the document d
Cp and Cn – positive and negative classes
score() – function that assigns positive or negative values
to terms
8/16
Document semantic orientation
• Semantic orientations of individual terms
are aggregated using a dictionary method
• This method uses two small sets
of manually identified
positive and negative adjectives,
which serve as seed sets
9/16
Document semantic orientation
• p( t|Cp ) and p( t|Cn ) – conditional
probabilities of the occurrence of the term t
in positive and negative class
• These probabilities may be approximated
by term occurrence frequencies
in the training set.
10 /16
Opinion prediction
• After document semantic orientation
and after removing the degree of subjectivity
• Algorithm for summarization of data
and prediction
11 /16
Improvements
• Defining more selective classes
• Assigning trust to credible users
• Using more social network data
to eliminate potential spams
12 /16
Advantages and practical uses
• It can help marketers
to evaluate the success of an ad campaign
or new product launch.
• Determine which versions of a product
or service are popular
and identify which people
will like or dislike product features
13/16
Disadvantages
• Can be very hard to determine the word class,
often depends of native language
• Requires strong machine learning algorithms
to solve classification problem
• Opinions are strongly relaying
on credibility of it’s users (social network)
14/16
Conclusion
• Using already existing data
• Fast growing technique,
follow grow of social media
15/16
Literature
• Milutinović, V., “The Best Method for Presentation of Research Results”,
IEEE TCCA
• R.F. Xu, , K.F. Wong, and Y.Q. WIA in NTCIR-7 MOAT Task, Japanese Weblog
Opinion Mining
• G. Wang, K. Araki,. Modifying SO , Opinion Mining by Using a Balancing
Factor and Detecting Neutral Expressions
• V. Hatzivassiloglou, , K.R. McKeown (1997). Predicting the semantic
orientation of adjectives. In Proceedings of the 35th Annual Meeting of the
Association for Computational Linguistics and the 8th Conference of the European
16/16
Questions?
Thank’s for your attention!