Most use of sentiment analysis in social media to date has been extremely limited. Analytics with dashboards full of traffic light symbols gloss only the most obvious features of social conversations, often obscuring the real reactions and trends which move opinion about a product or a company. In this presentation, we discuss the causes, both technical and human, behind the failure of early sentiment approaches. We will introduce the technologies and practices for advanced conversation analytics, and show how understanding tone and context for commentary provides a far more accurate analytical frame for decision-making around social media.
Improving Your Literature Reviews with NVivo 10 for Windows
Beyond Sentiment Hype: Conversation Context for Accurate Discovery
1. Beyond
Sen)ment
Hype:
Conversa)on
Context
for
Accurate
Discovery
Hadley
Reynolds
NextEra
Research
2. Agenda
• Where
we
are
now
–
market
drivers
&
technology
dynamics
• The
Sen)ment
Bubble
considered
• Differen)a)ng
levels
of
analysis
• Prac)cal
dimensions
of
analysis
and
examples
• Discussion
7. Market
Drivers
for
Sen)ment
Analysis
Addi$onal
Web
2.0
Content:
Blogs
Discussion
Forums
Amazon
(Yelp,
Trip
Advisor
etc.)
Reviews
User
Generated
RaAngs
Data
“Like”
Google+
And
more,
much
more…
13. Challenges
for
Sen)ment
Analysis
• Level
of
analysis
• Timeframes
for
analysis
• Rela)ve
sophis)ca)on
of
analysis
14. Level
of
Analysis
• Corpus
(Do
the
bloggers
like
us?)
• Document
(Does
this
author
like
us?)
15. Document
Sen)ment
Math
Posi)ve
document
=
4
points
or
above
Nega)ve
document
=
-‐2
points
or
below
Neutral
document
=
-‐2
through
+3
good
Value
Score
great
Term
good
2
2
o.k.
great
3
3
o.k.
1
1
disappointed
-‐4
-‐4
Total:
+2
disappointed
Neutral
Document
16. Document
Sen)ment
Math
Posi)ve
document
=
4
points
or
above
Nega)ve
document
=
-‐2
points
or
below
Product
A
Product
B
Neutral
document
=
-‐2
through
+3
good
ok
good
Value
Score
great
Term
ok
Product
A
good
2
2
o.k.
Product
A
great
3
3
Product
A
o.k.
1
1
Product
A
-‐4
-‐4
disappointed
Product
B
good
1
1
disappointed
Product
B
ok
1
2
disappointed
Product
B
-‐4
-‐8
disappointed
disappointed
Nega)ve
Document
Total:
-‐3
17. Level
of
Analysis
• Corpus
(Do
the
bloggers
like
us?)
• Document
(Does
this
author
like
us?)
• Sentence
(What
is
this
person’s
comment?)
• En)ty/A`ribute
(What
is
it
about
us
that
she
likes
or
doesn’t
like?)
18. En)ty-‐level
Analysis
Sources
Person
Opinion
Target
En)ty
(Feature)
(Profile)
Person
(Emo)on)
Opinion
(Feature)
Target
En)ty
(Feature)
(Social
Network)
20. Sophis)ca)on
of
Analysis
• Keyword-‐based
sen)ment
techniques
– Sen)ment
terms:
elusive,
ambiguous,
in
flux
– Sen)ment
lexicons:
incomplete,
non-‐specific,
inflexible
– Unable
to
understand
context
surrounding
an
expression
or
the
people
contribu)ng
– Unable
to
understand
connec)ons
among
related
en))es
and
a`ributes
and
people
– Unable
to
gauge
quality
of
source
materials
21. Sophis)ca)on
of
Analysis
• Seman)c-‐based
sen)ment
techniques
– Sen)ment
terms
>>
incorporate
related
expressions,
fuzzy
logic
-‐
NLP
– Sen)ment
lexicons
>>
domain
ontologies
(available
or
buildable)
provide
analy)cal
context
– Able
to
understand
context
surrounding
an
expression
or
the
people
contribu)ng
-‐
machine
learning
&
other
techniques
– Able
to
understand
connec)ons
among
related
en))es
and
a`ributes
and
people
-‐
triples,
event
extrac)on
22. Dimensions
of
Analysis
• Ontologies
around
opinion
objects
• Iden)fica)on
and
qualifica)on
of
en))es
&
a`ributes
&
rela)onships
• Emo)onal
content
of
expression(s)
• Quality
gauge
of
sources
• Profiles
of
individual
commenters
• Roles/interac)ons/sociology
of
commenters
and
their
affilia)ons
• Timeframe
for
expressions
and
responses
23. Beyond
+/-‐:
Ontology-‐based
analy)cs
Same
Ontology
breakdown
Same
Scale:
Expressed
Opinions
Higher
values
for
cardiovascular
diseases
with
Avas)n
Source:
BuzzStory
25. Quality
of
Content
Sources
topix.com
cancergrace.org
• Quality:
4.48
• Quality:
16.78
"I
know
of
one
method
that
"As
shown
above,
a
total
of
would
be
really
scary
and
362
pa)ents
who
hadn't
graphic
that
would
work
progressed
aser
first
line
towards
gepng
people
to
chemo/Avas)n
were
stop
pollu)ng
my
sea
breeze
randomized
to
either
of
the
environment.
two
maintenance
therapy
What
I
wonna
know
is
they
arms,
and
the
combina)on
keep
pupng
down
smokers
arm
showed
a
significantly
and
blaming
us
for
longer
progression-‐free
evrything.”
survival
(PFS)
coun)ng
from
the
beginning
of
all
treatment,
at
10.”
26. Affilia)on
Network
–
Map
of
Affilia)ons
of
People
&
Topics
Supplements
Tobacco
Addic)on
Prostate
Cancer
Breast
Cancer
Co-‐Morbidi)es
Thyroid
Disease
Biomarkers
Lung
Cancer
Targeted
Therapies
Chemotherapy
H&N
Cancer
Source:
BuzzStory
27. Sociology
of
Affilia)ons
&
Topic
Groupings
Co-‐Morbidi)es
Tobacco
Addic)on
Other
Types
of
Cancer
Supplements
Misc.
Side-‐Effects
Misc.
Side-‐Effects
Biomarkers
Source:
BuzzStory
29. Challenges
Remain
“The
service
at
Reynards
is,
in
general,
friendly
and
loose.
Though
they
couldn’t
find
a
reserva)on
for
four
one
Friday
night,
they
compensated
with
so
much
warmth
and
comped
wine
that
all
was
forgiven.
In
some
ways,
Reynards
offers
what
one
wishes
a
dining
experience
in
Manha`an
would
be:
kindness
instead
of
aptude,
inoffensive
prices,
glorious
food,
and
aesthe)c
variety—the
clientele
is
split
roughly
in
half
between
the
stylish
and
the
schlumpy.”
The
New
Yorker,
September
24,
2012
30. Resources
• Bing
Liu,
Sen$ment
Analysis
and
Opinion
Mining,
Morgan
&
Claypool,
2012
• Bo
Pang
and
Lillian
Lee,
Opinion
Mining
and
Sen$ment
Analysis,
(Founda$ons
and
Trends
in
Informa$on
Retrieval),
Now
Publishers,
2008
• Sen)ment
Analysis
Symposium,
San
Francisco,
CA,
October
30,
2012