<s>
Personalized	B-Application
search	I-Application
refers	O
to	O
web	B-Application
search	I-Application
experiences	O
that	O
are	O
tailored	O
specifically	O
to	O
an	O
individual	O
's	O
interests	O
by	O
incorporating	O
information	O
about	O
the	O
individual	O
beyond	O
the	O
specific	O
query	O
provided	O
.	O
</s>
<s>
There	O
are	O
two	O
general	O
approaches	O
to	O
personalizing	O
search	B-Application
results	I-Application
,	O
involving	O
modifying	O
the	O
user	O
's	O
query	O
and	O
re-ranking	O
search	B-Application
results	I-Application
.	O
</s>
<s>
Google	B-Application
introduced	O
personalized	B-Application
search	I-Application
in	O
2004	O
and	O
it	O
was	O
implemented	O
in	O
2005	O
to	O
Google	B-Application
search	O
.	O
</s>
<s>
Google	B-Application
has	O
personalized	B-Application
search	I-Application
implemented	O
for	O
all	O
users	O
,	O
not	O
only	O
those	O
with	O
a	O
Google	B-Application
account	O
.	O
</s>
<s>
There	O
is	O
not	O
much	O
information	O
on	O
how	O
exactly	O
Google	B-Application
personalizes	O
their	O
searches	O
;	O
however	O
,	O
it	O
is	O
believed	O
that	O
they	O
use	O
user	O
language	O
,	O
location	O
,	O
and	O
web	O
history	O
.	O
</s>
<s>
Early	O
search	B-Application
engines	I-Application
,	O
like	O
Google	B-Application
and	O
AltaVista	O
,	O
found	O
results	O
based	O
only	O
on	O
key	O
words	O
.	O
</s>
<s>
Personalized	B-Application
search	I-Application
,	O
as	O
pioneered	O
by	O
Google	B-Application
,	O
has	O
become	O
far	O
more	O
complex	O
with	O
the	O
goal	O
to	O
"	O
understand	O
exactly	O
what	O
you	O
mean	O
and	O
give	O
you	O
exactly	O
what	O
you	O
want.	O
"	O
</s>
<s>
Using	O
mathematical	O
algorithms	O
,	O
search	B-Application
engines	I-Application
are	O
now	O
able	O
to	O
return	O
results	O
based	O
on	O
the	O
number	O
of	O
links	O
to	O
and	O
from	O
sites	O
;	O
the	O
more	O
links	O
a	O
site	O
has	O
,	O
the	O
higher	O
it	O
is	O
placed	O
on	O
the	O
page	O
.	O
</s>
<s>
Search	B-Application
engines	I-Application
have	O
two	O
degrees	O
of	O
expertise	O
:	O
the	O
shallow	O
expert	O
and	O
the	O
deep	O
expert	O
.	O
</s>
<s>
If	O
a	O
person	O
knows	O
what	O
he	O
or	O
she	O
wants	O
then	O
the	O
search	B-Application
engine	I-Application
will	O
act	O
as	O
a	O
shallow	O
expert	O
and	O
simply	O
locate	O
that	O
information	O
.	O
</s>
<s>
But	O
search	B-Application
engines	I-Application
are	O
also	O
capable	O
of	O
deep	O
expertise	O
in	O
that	O
they	O
rank	O
results	O
indicating	O
that	O
those	O
near	O
the	O
top	O
are	O
more	O
relevant	O
to	O
a	O
user	O
's	O
wants	O
than	O
those	O
below	O
.	O
</s>
<s>
While	O
many	O
search	B-Application
engines	I-Application
take	O
advantage	O
of	O
information	O
about	O
people	O
in	O
general	O
,	O
or	O
about	O
specific	O
groups	O
of	O
people	O
,	O
personalized	B-Application
search	I-Application
depends	O
on	O
a	O
user	O
profile	O
that	O
is	O
unique	O
to	O
the	O
individual	O
.	O
</s>
<s>
Research	O
systems	O
that	O
personalize	O
search	B-Application
results	I-Application
model	O
their	O
users	O
in	O
different	O
ways	O
.	O
</s>
<s>
There	O
are	O
several	O
publicly	O
available	O
systems	O
for	O
personalizing	O
Web	B-Application
search	I-Application
results	O
(	O
e.g.	O
,	O
Google	B-Application
Personalized	I-Application
Search	I-Application
and	O
Bing	B-Application
's	O
search	O
result	O
personalization	O
)	O
.	O
</s>
<s>
One	O
technique	O
Google	B-Application
uses	O
to	O
personalize	O
searches	O
for	O
its	O
users	O
is	O
to	O
track	O
log	O
in	O
time	O
and	O
if	O
the	O
user	O
has	O
enabled	O
web	O
history	O
in	O
his	O
browser	O
.	O
</s>
<s>
If	O
a	O
user	O
accesses	O
the	O
same	O
site	O
through	O
a	O
search	O
result	O
from	O
Google	B-Application
many	O
times	O
,	O
it	O
believes	O
that	O
they	O
like	O
that	O
page	O
.	O
</s>
<s>
So	O
when	O
users	O
carry	O
out	O
certain	O
searches	O
,	O
Google	B-Application
's	I-Application
personalized	B-Application
search	I-Application
algorithm	O
gives	O
the	O
page	O
a	O
boost	O
,	O
moving	O
it	O
up	O
through	O
the	O
ranks	O
.	O
</s>
<s>
Even	O
if	O
a	O
user	O
is	O
signed	O
out	O
,	O
Google	B-Application
may	O
personalize	O
their	O
results	O
because	O
it	O
keeps	O
a	O
180-day	O
record	O
of	O
what	O
a	O
particular	O
web	O
browser	O
has	O
searched	O
for	O
,	O
linked	O
to	O
a	O
cookie	O
in	O
that	O
browser	O
.	O
</s>
<s>
In	O
search	B-Application
engines	I-Application
on	O
social	O
networking	O
platforms	O
like	O
Facebook	B-Application
or	O
LinkedIn	B-Application
,	O
personalization	O
could	O
be	O
achieved	O
by	O
exploiting	O
homophily	O
between	O
searchers	O
and	O
results	O
.	O
</s>
<s>
In	O
Job	O
search	O
,	O
searchers	O
are	O
usually	O
interested	O
in	O
jobs	O
at	O
similar	O
companies	O
,	O
jobs	O
at	B-Application
nearby	I-Application
locations	I-Application
and	O
jobs	O
requiring	O
expertise	O
similar	O
to	O
their	O
own	O
.	O
</s>
<s>
In	O
order	O
to	O
better	O
understand	O
how	O
personalized	B-Application
search	I-Application
results	O
are	O
being	O
presented	O
to	O
the	O
users	O
,	O
a	O
group	O
of	O
researchers	O
at	O
Northeastern	O
University	O
compared	O
an	O
aggregate	O
set	O
of	O
searches	O
from	O
logged	O
in	O
users	O
against	O
a	O
control	O
group	O
.	O
</s>
<s>
Of	O
various	O
factors	O
tested	O
,	O
the	O
two	O
that	O
had	O
measurable	O
impact	O
were	O
being	O
logged	O
in	O
with	O
a	O
Google	B-Application
account	O
and	O
the	O
IP	B-Protocol
address	I-Protocol
of	O
the	O
searching	O
users	O
.	O
</s>
<s>
So	O
,	O
for	O
example	O
,	O
if	O
a	O
user	O
searched	O
for	O
"	O
used	O
car	O
sales	O
"	O
,	O
Google	B-Application
may	O
produce	O
results	O
of	O
local	O
car	O
dealerships	O
in	O
their	O
area	O
.	O
</s>
<s>
Several	O
concerns	O
have	O
been	O
brought	O
up	O
regarding	O
personalized	B-Application
search	I-Application
.	O
</s>
<s>
It	O
decreases	O
the	O
likelihood	O
of	O
finding	O
new	O
information	O
by	O
biasing	B-Application
search	I-Application
results	I-Application
towards	O
what	O
the	O
user	O
has	O
already	O
found	O
.	O
</s>
<s>
It	O
introduces	O
potential	O
privacy	O
problems	O
in	O
which	O
a	O
user	O
may	O
not	O
be	O
aware	O
that	O
their	O
search	B-Application
results	I-Application
are	O
personalized	O
for	O
them	O
,	O
and	O
wonder	O
why	O
the	O
things	O
that	O
they	O
are	O
interested	O
in	O
have	O
become	O
so	O
relevant	O
.	O
</s>
<s>
Such	O
a	O
problem	O
has	O
been	O
coined	O
as	O
the	O
"	O
filter	B-Algorithm
bubble	I-Algorithm
"	O
by	O
author	O
Eli	O
Pariser	O
.	O
</s>
<s>
This	O
can	O
isolate	O
users	O
in	O
their	O
own	O
worlds	O
or	O
"	O
filter	B-Algorithm
bubbles	I-Algorithm
"	O
where	O
they	O
only	O
see	O
information	O
that	O
they	O
want	O
to	O
,	O
such	O
a	O
consequence	O
of	O
"	O
The	O
Friendly	O
World	O
Syndrome	O
"	O
.	O
</s>
<s>
The	O
personalization	O
method	O
makes	O
it	O
very	O
easy	O
to	O
understand	O
how	O
the	B-Algorithm
filter	I-Algorithm
bubble	I-Algorithm
is	O
created	O
.	O
</s>
<s>
Filter	B-Algorithm
bubbles	I-Algorithm
have	O
become	O
more	O
frequent	O
in	O
search	B-Application
results	I-Application
and	O
are	O
envisaged	O
as	O
disruptions	O
to	O
information	O
flow	O
in	O
online	O
more	O
specifically	O
social	O
media	O
.	O
</s>
<s>
An	O
area	O
of	O
particular	O
concern	O
to	O
some	O
parts	O
of	O
the	O
world	O
is	O
the	O
use	O
of	O
personalized	B-Application
search	I-Application
as	O
a	O
form	O
of	O
control	O
over	O
the	O
people	O
utilizing	O
the	O
search	O
by	O
only	O
giving	O
them	O
particular	O
information	O
(	O
selective	O
exposure	O
)	O
.	O
</s>
<s>
While	O
total	O
control	O
by	O
a	O
particular	O
government	O
just	O
from	O
personalized	B-Application
search	I-Application
is	O
a	O
stretch	O
,	O
control	O
of	O
the	O
information	O
readily	O
available	O
from	O
searches	O
can	O
easily	O
be	O
controlled	O
by	O
the	O
richest	O
corporations	O
.	O
</s>
<s>
The	O
biggest	O
example	O
of	O
a	O
corporation	O
controlling	O
the	O
information	O
is	O
Google	B-Application
.	I-Application
</s>
<s>
Google	B-Application
is	O
not	O
only	O
feeding	O
you	O
the	O
information	O
they	O
want	O
but	O
they	O
are	O
at	O
times	O
using	O
your	O
personalized	B-Application
search	I-Application
to	O
gear	O
you	O
towards	O
their	O
own	O
companies	O
or	O
affiliates	O
.	O
</s>
<s>
This	O
has	O
led	O
to	O
a	O
complete	O
control	O
of	O
various	O
parts	O
of	O
the	O
web	O
and	O
a	O
pushing	O
out	O
of	O
their	O
competitors	O
such	O
as	O
how	O
Google	B-Application
Maps	O
took	O
a	O
major	O
control	O
over	O
the	O
online	O
map	O
and	O
direction	O
industry	O
with	O
MapQuest	O
and	O
others	O
forced	O
to	O
take	O
a	O
backseat	O
.	O
</s>
<s>
Many	O
search	B-Application
engines	I-Application
use	O
concept-based	O
user	O
profiling	O
strategies	O
that	O
derive	O
only	O
topics	O
that	O
users	O
are	O
highly	O
interested	O
in	O
but	O
for	O
best	O
results	O
,	O
according	O
to	O
researchers	O
Wai-Tin	O
and	O
Dik	O
Lun	O
,	O
both	O
positive	O
and	O
negative	O
preferences	O
should	O
be	O
considered	O
.	O
</s>
<s>
For	O
example	O
,	O
typing	O
in	O
'	O
apple	O
 '	O
could	O
refer	O
to	O
either	O
the	O
fruit	O
or	O
the	O
Macintosh	B-Device
computer	O
and	O
providing	O
both	O
preferences	O
aids	O
search	B-Application
engines	I-Application
 '	O
ability	O
to	O
learn	O
which	O
apple	O
the	O
user	O
is	O
really	O
looking	O
for	O
based	O
on	O
the	O
links	O
clicked	O
.	O
</s>
<s>
One	O
concept-strategy	O
the	O
researchers	O
came	O
up	O
with	O
to	O
improve	O
personalized	B-Application
search	I-Application
and	O
yield	O
both	O
positive	O
and	O
negative	O
preferences	O
is	O
the	O
click-based	O
method	O
.	O
</s>
<s>
The	O
feature	O
also	O
has	O
profound	O
effects	O
on	O
the	O
search	B-Application
engine	I-Application
optimization	O
industry	O
,	O
due	O
to	O
the	O
fact	O
that	O
search	B-Application
results	I-Application
will	O
no	O
longer	O
be	O
ranked	O
the	O
same	O
way	O
for	O
every	O
user	O
.	O
</s>
<s>
An	O
example	O
of	O
this	O
is	O
found	O
in	O
Eli	O
Pariser	O
's	O
,	O
The	B-Algorithm
Filter	I-Algorithm
Bubble	I-Algorithm
,	O
where	O
he	O
had	O
two	O
friends	O
type	O
in	O
"	O
BP	O
"	O
into	O
Google	B-Application
's	I-Application
search	O
bar	O
.	O
</s>
<s>
The	O
aspect	O
of	O
information	O
overload	O
is	O
also	O
prevalent	O
when	O
using	O
search	B-Application
engine	I-Application
optimization	O
.	O
</s>
<s>
For	O
instance	O
,	O
Google	B-Application
uses	O
various	O
‘‘	O
signals’’	O
in	O
order	O
to	O
personalize	O
searches	O
including	O
location	O
,	O
previous	O
search	O
keywords	O
and	O
recently	O
contacts	O
in	O
a	O
user	O
’s	O
social	O
network	O
while	O
on	O
the	O
other	O
hand	O
,	O
Facebook	B-Application
registers	O
the	O
user	O
’s	O
interactions	O
with	O
other	O
users	O
,	O
the	O
so-called	O
‘‘’’	O
.	O
</s>
<s>
The	B-Algorithm
filter	I-Algorithm
bubble	I-Algorithm
has	O
made	O
a	O
heavy	O
effect	O
on	O
the	O
search	O
for	O
information	O
of	O
health	O
.	O
</s>
<s>
With	O
the	O
influence	O
of	O
search	B-Application
results	I-Application
based	O
upon	O
search	O
history	O
,	O
social	O
network	O
,	O
personal	O
preference	O
and	O
other	O
aspects	O
,	O
misinformation	O
has	O
been	O
a	O
large	O
contributor	O
in	O
the	O
drop	O
of	O
vaccination	O
rate	O
.	O
</s>
<s>
Some	O
have	O
noted	O
that	O
personalized	B-Application
search	I-Application
results	I-Application
not	O
only	O
serve	O
to	O
customize	O
a	O
user	O
's	O
search	B-Application
results	I-Application
,	O
but	O
also	O
advertisements	O
.	O
</s>
<s>
An	O
important	O
example	O
of	O
search	O
personalization	O
is	O
Google	B-Application
.	I-Application
</s>
<s>
There	O
are	O
a	O
host	O
of	O
Google	B-Application
applications	O
,	O
all	O
of	O
which	O
can	O
be	O
personalized	O
and	O
integrated	O
with	O
the	O
help	O
of	O
a	O
Google	B-Application
account	O
.	O
</s>
<s>
However	O
,	O
one	O
is	O
almost	O
deprived	O
of	O
a	O
choice	O
,	O
since	O
so	O
many	O
useful	O
Google	B-Application
products	O
are	O
only	O
accessible	O
if	O
one	O
has	O
a	O
Google	B-Application
account	O
.	O
</s>
<s>
The	B-Application
Google	I-Application
Dashboard	O
,	O
introduced	O
in	O
2009	O
,	O
covers	O
more	O
than	O
20	O
products	O
and	O
services	O
,	O
including	O
Gmail	O
,	O
Calendar	O
,	O
Docs	O
,	O
YouTube	O
,	O
etc	O
.	O
</s>
<s>
The	O
free	O
Google	B-Application
Custom	O
Search	O
is	O
available	O
for	O
individuals	O
and	O
big	O
companies	O
alike	O
,	O
providing	O
the	O
Search	O
facility	O
for	O
individual	O
websites	O
and	O
powering	O
corporate	O
sites	O
such	O
as	O
that	O
of	O
the	O
New	O
York	O
Times	O
.	O
</s>
<s>
The	O
high	O
level	O
of	O
personalization	O
that	O
was	O
available	O
with	O
Google	B-Application
played	O
a	O
significant	O
part	O
in	O
helping	O
remain	O
the	O
world	O
's	O
most	O
favorite	O
search	B-Application
engine	I-Application
.	O
</s>
<s>
One	O
example	O
of	O
Google	B-Application
's	I-Application
ability	O
to	O
personalize	O
searches	O
is	O
in	O
its	O
use	O
of	O
Google	B-Application
News	O
.	O
</s>
<s>
Google	B-Application
has	O
geared	O
its	O
news	O
to	O
show	O
everyone	O
a	O
few	O
similar	O
articles	O
that	O
can	O
be	O
deemed	O
interesting	O
,	O
but	O
as	O
soon	O
as	O
the	O
user	O
scrolls	O
down	O
,	O
it	O
can	O
be	O
seen	O
that	O
the	O
news	O
articles	O
begin	O
to	O
differ	O
.	O
</s>
<s>
Google	B-Application
takes	O
into	O
account	O
past	O
searches	O
as	O
well	O
as	O
the	O
location	O
of	O
the	O
user	O
to	O
make	O
sure	O
that	O
local	O
news	O
gets	O
to	O
them	O
first	O
.	O
</s>
<s>
This	O
can	O
create	O
the	O
"	O
filter	B-Algorithm
bubble	I-Algorithm
"	O
as	O
described	O
earlier	O
.	O
</s>
<s>
Google	B-Application
provides	O
a	O
host	O
of	O
services	O
to	O
people	O
,	O
and	O
many	O
of	O
these	O
services	O
do	O
not	O
require	O
information	O
to	O
be	O
collected	O
about	O
a	O
person	O
to	O
be	O
customizable	O
.	O
</s>
<s>
As	O
people	O
reap	O
the	O
rewards	O
of	O
convenience	O
from	O
customizing	O
their	O
other	O
Google	B-Application
services	O
,	O
they	O
desire	O
better	O
search	B-Application
results	I-Application
,	O
even	O
if	O
it	O
comes	O
at	O
the	O
expense	O
of	O
private	O
information	O
.	O
</s>
<s>
Where	O
to	O
draw	O
the	O
line	O
between	O
the	O
information	O
versus	O
search	B-Application
results	I-Application
tradeoff	O
is	O
new	O
territory	O
and	O
Google	B-Application
gets	O
to	O
make	O
that	O
decision	O
.	O
</s>
<s>
Until	O
people	O
get	O
the	O
power	O
to	O
control	O
the	O
information	O
that	O
is	O
being	O
collected	O
about	O
them	O
,	O
Google	B-Application
is	O
not	O
truly	O
protecting	O
privacy	O
.	O
</s>
<s>
Google	B-Application
's	I-Application
popularity	O
as	O
a	O
search	B-Application
engine	I-Application
and	O
Internet	O
browser	O
has	O
allowed	O
it	O
to	O
gain	O
a	O
lot	O
of	O
power	O
.	O
</s>
<s>
Google	B-Application
can	O
use	O
multiple	O
methods	O
of	O
personalization	O
such	O
as	O
traditional	O
,	O
social	O
,	O
geographic	O
,	O
IP	B-Protocol
address	I-Protocol
,	O
browser	O
,	O
cookies	O
,	O
time	O
of	O
day	O
,	O
year	O
,	O
behavioral	O
,	O
query	O
history	O
,	O
bookmarks	O
,	O
and	O
more	O
.	O
</s>
<s>
Although	O
having	O
Google	B-Application
personalize	O
search	B-Application
results	I-Application
based	O
on	O
what	O
users	O
searched	O
previously	O
may	O
have	O
its	O
benefits	O
,	O
there	O
are	O
negatives	O
that	O
come	O
with	O
it	O
.	O
</s>
<s>
With	O
the	O
power	O
from	O
this	O
information	O
,	O
Google	B-Application
has	O
chosen	O
to	O
enter	O
other	O
sectors	O
it	O
owned	O
,	O
such	O
as	O
videos	O
,	O
document	O
sharing	O
,	O
shopping	O
,	O
maps	O
,	O
and	O
many	O
more	O
.	O
</s>
<s>
Google	B-Application
has	O
done	O
this	O
by	O
steering	O
searchers	O
to	O
their	O
own	O
services	O
offered	O
as	O
opposed	O
to	O
others	O
such	O
as	O
MapQuest	O
.	O
</s>
<s>
Using	O
search	O
personalization	O
,	O
Google	B-Application
has	O
doubled	O
its	O
video	O
market	O
share	O
to	O
about	O
eighty	O
percent	O
.	O
</s>
<s>
The	O
legal	O
definition	O
of	O
a	O
monopoly	B-Application
is	O
when	O
a	O
firm	O
gains	O
control	O
of	O
seventy	O
to	O
eighty	O
percent	O
of	O
the	O
market	O
.	O
</s>
<s>
Google	B-Application
has	O
reinforced	O
this	O
monopoly	B-Application
by	O
creating	O
significant	O
barriers	O
of	O
entry	O
such	O
as	O
manipulating	O
search	B-Application
results	I-Application
to	O
show	O
their	O
own	O
services	O
.	O
</s>
<s>
This	O
can	O
be	O
clearly	O
seen	O
with	O
Google	B-Application
Maps	O
being	O
the	O
first	O
thing	O
displayed	O
in	O
most	O
searches	O
.	O
</s>
<s>
It	O
becomes	O
apparent	O
that	O
the	O
decline	O
of	O
these	O
companies	O
has	O
come	O
because	O
of	O
Google	B-Application
's	I-Application
increase	O
in	O
market	O
share	O
from	O
43%	O
in	O
2007	O
to	O
about	O
55%	O
in	O
2009	O
.	O
</s>
<s>
It	O
can	O
be	O
said	O
that	O
Google	B-Application
is	O
more	O
dominant	O
because	O
they	O
provide	O
better	O
services	O
.	O
</s>
<s>
The	O
graph	O
for	O
product	O
search	O
is	O
evidence	O
enough	O
for	O
Google	B-Application
's	I-Application
influence	O
because	O
their	O
numbers	O
went	O
from	O
1.3	O
million	O
unique	O
visitors	O
to	O
11.9	O
unique	O
visitors	O
in	O
one	O
month	O
.	O
</s>
<s>
The	O
first	O
is	O
that	O
Google	B-Application
's	I-Application
market	O
share	O
has	O
a	O
direct	O
inverse	O
relationship	O
to	O
the	O
market	O
share	O
of	O
the	O
leading	O
competitors	O
.	O
</s>
<s>
The	O
second	O
is	O
that	O
this	O
directly	O
inverse	O
relationship	O
began	O
around	O
2007	O
,	O
which	O
is	O
around	O
the	O
time	O
that	O
Google	B-Application
began	O
to	O
use	O
its	O
""	O
method	O
.	O
</s>
<s>
One	O
of	O
the	O
most	O
critical	O
benefits	O
personalized	B-Application
search	I-Application
has	O
is	O
to	O
improve	O
the	O
quality	O
of	O
decisions	O
consumers	O
make	O
.	O
</s>
<s>
Two	O
studies	O
examined	O
the	O
effects	O
of	O
personalized	O
screening	O
and	O
ordering	O
tools	O
,	O
and	O
the	O
results	O
show	O
a	O
positive	O
correlation	O
between	O
personalized	B-Application
search	I-Application
and	O
the	O
quality	O
of	O
consumers	O
 '	O
decisions	O
.	O
</s>
<s>
Dellaert	O
from	O
Maastricht	O
University	O
mainly	O
focused	O
on	O
recommendation	B-Application
systems	I-Application
.	O
</s>
<s>
Both	O
studies	O
concluded	O
that	O
a	O
personalized	B-Application
search	I-Application
and	O
recommendation	B-Application
system	I-Application
significantly	O
improved	O
consumers	O
 '	O
decision	O
quality	O
and	O
reduced	O
the	O
number	O
of	O
products	O
inspected	O
.	O
</s>
<s>
On	O
the	O
same	O
note	O
the	O
use	O
of	O
the	O
use	O
of	O
filter	B-Algorithm
bubbles	I-Algorithm
in	O
personalized	B-Application
search	I-Application
has	O
also	O
led	O
to	O
several	O
benefits	O
to	O
the	O
users	O
.	O
</s>
<s>
For	O
instance	O
filter	B-Algorithm
bubbles	I-Algorithm
have	O
the	O
potential	O
of	O
enhancing	O
opinion	O
diversity	O
by	O
allowing	O
like-minded	O
citizens	O
to	O
come	O
together	O
and	O
reinforce	O
their	O
beliefs	O
.	O
</s>
<s>
This	O
also	O
helps	O
in	O
protecting	O
users	O
from	O
fake	O
and	O
extremist	O
content	O
by	O
enclosing	O
them	O
in	O
bubbles	B-Algorithm
of	O
reliable	O
and	O
verifiable	O
information	O
.	O
</s>
<s>
Filter	B-Algorithm
bubbles	I-Algorithm
can	O
be	O
an	O
important	O
element	O
of	O
information	O
freedom	O
by	O
providing	O
users	O
more	O
choice	O
.	O
</s>
<s>
Personalized	B-Application
search	I-Application
has	O
also	O
proved	O
to	O
work	O
on	O
the	O
benefit	O
of	O
the	O
user	O
in	O
the	O
sense	O
that	O
they	O
improve	O
the	O
information	O
search	B-Application
results	I-Application
.	O
</s>
<s>
Personalized	B-Application
search	I-Application
tailors	O
search	O
result	O
to	O
the	O
needs	O
of	O
the	O
user	O
in	O
the	O
sense	O
that	O
it	O
matches	O
what	O
the	O
user	O
wants	O
with	O
past	O
search	O
history	O
.	O
</s>
<s>
For	O
instance	O
,	O
in	O
Google	B-Application
,	O
the	O
search	O
history	O
of	O
user	O
is	O
kept	O
and	O
matched	O
with	O
the	O
user	O
query	O
in	O
the	O
user	O
's	O
next	O
searches	O
.	O
</s>
<s>
Google	B-Application
achieves	O
this	O
through	O
three	O
important	O
techniques	O
.	O
</s>
<s>
Personalized	B-Application
search	I-Application
gains	O
popularity	O
because	O
of	O
the	O
demand	O
for	O
more	O
relevant	O
information	O
and	O
the	O
fact	O
that	O
most	O
people	O
could	O
really	O
use	O
some	O
personal	O
information	O
such	O
as	O
personalized	B-Application
search	I-Application
gains	O
.	O
</s>
<s>
Research	O
has	O
indicated	O
low	O
success	O
rates	O
among	O
major	O
search	B-Application
engines	I-Application
in	O
providing	O
relevant	O
results	O
;	O
in	O
52%	O
of	O
20,000	O
queries	O
,	O
searchers	O
did	O
not	O
find	O
any	O
relevant	O
results	O
within	O
the	O
documents	O
that	O
Google	B-Application
returned	O
.	O
</s>
<s>
Personalized	B-Application
search	I-Application
can	O
improve	O
search	O
quality	O
significantly	O
and	O
there	O
are	O
mainly	O
two	O
ways	O
to	O
achieve	O
this	O
goal	O
.	O
</s>
<s>
There	O
is	O
another	O
way	O
to	O
personalize	O
search	B-Application
results	I-Application
.	O
</s>
<s>
In	O
Bracha	O
Shapira	O
and	O
Boaz	O
Zabar	O
's	O
"	O
Personalized	B-Application
Search	I-Application
:	O
Integrating	O
Collaboration	O
and	O
Social	O
Networks	O
"	O
,	O
Shapira	O
and	O
Zabar	O
focused	O
on	O
a	O
model	O
that	O
utilizes	O
a	O
recommendation	B-Application
system	I-Application
.	O
</s>
<s>
The	O
authors	O
examined	O
keyword	O
search	O
,	O
the	O
recommendation	B-Application
system	I-Application
,	O
and	O
the	O
recommendation	B-Application
system	I-Application
with	O
social	O
network	O
working	O
separately	O
and	O
compares	O
the	O
results	O
in	O
terms	O
of	O
search	O
quality	O
.	O
</s>
<s>
The	O
results	O
show	O
that	O
a	O
personalized	B-Application
search	I-Application
engine	I-Application
with	O
the	O
recommendation	B-Application
system	I-Application
produces	O
better	O
quality	O
results	O
than	O
the	O
standard	O
search	B-Application
engine	I-Application
,	O
and	O
that	O
the	O
recommendation	B-Application
system	I-Application
with	O
social	O
network	O
even	O
improves	O
more	O
.	O
</s>
<s>
Recent	O
paper	O
“”	O
shows	O
that	O
a	O
new	O
embedding	O
model	O
for	O
search	O
personalization	O
,	O
where	O
users	O
are	O
embedded	O
on	O
a	O
topical	O
interest	O
space	O
,	O
produces	O
better	O
search	B-Application
results	I-Application
than	O
strong	O
learning-to-rank	O
models	O
.	O
</s>
<s>
The	O
foundation	O
of	O
this	O
argument	O
against	O
its	O
use	O
is	O
because	O
it	O
confines	O
internet	O
users	O
 '	O
search	B-Application
engine	I-Application
results	O
to	O
material	O
that	O
aligns	O
with	O
the	O
users	O
 '	O
interests	O
and	O
history	O
.	O
</s>
<s>
Search	O
personalization	O
takes	O
the	O
objectivity	O
out	O
of	O
the	O
search	B-Application
engine	I-Application
and	O
undermines	O
the	O
engine	O
.	O
</s>
<s>
If	O
a	O
user	O
has	O
a	O
particular	O
set	O
of	O
interests	O
or	O
internet	O
history	O
and	O
uses	O
the	O
web	O
to	O
research	O
a	O
controversial	O
issue	O
,	O
the	O
user	O
's	O
search	B-Application
results	I-Application
will	O
reflect	O
that	O
.	O
</s>
<s>
A	O
study	O
done	O
on	O
search	O
personalization	O
and	O
its	O
effects	O
on	O
search	B-Application
results	I-Application
in	O
Google	B-Application
News	O
resulted	O
in	O
different	O
orders	O
of	O
news	O
stories	O
being	O
generated	O
by	O
different	O
users	O
,	O
even	O
though	O
each	O
user	O
entered	O
the	O
same	O
search	O
query	O
.	O
</s>
<s>
Another	O
disadvantage	O
of	O
search	O
personalization	O
is	O
that	O
internet	O
companies	O
such	O
as	O
Google	B-Application
are	O
gathering	O
and	O
potentially	O
selling	O
their	O
users	O
 '	O
internet	O
interests	O
and	O
histories	O
to	O
other	O
companies	O
.	O
</s>
<s>
E	O
.	O
Pariser	O
,	O
author	O
of	O
The	B-Algorithm
Filter	I-Algorithm
Bubble	I-Algorithm
,	O
explains	O
how	O
there	O
are	O
differences	O
that	O
search	O
personalization	O
has	O
on	O
both	O
Facebook	B-Application
and	O
Google	B-Application
.	I-Application
</s>
<s>
Facebook	B-Application
implements	O
personalization	O
when	O
it	O
comes	O
to	O
the	O
amount	O
of	O
things	O
people	O
share	O
and	O
what	O
pages	O
they	O
"	O
like	O
"	O
.	O
</s>
<s>
An	O
individual	O
's	O
social	O
interactions	O
,	O
whose	O
profile	O
they	O
visit	O
the	O
most	O
,	O
who	O
they	O
message	O
or	O
chat	O
with	O
are	O
all	O
indicators	O
that	O
are	O
used	O
when	O
Facebook	B-Application
uses	O
personalization	O
.	O
</s>
<s>
Rather	O
than	O
what	O
people	O
share	O
being	O
an	O
indicator	O
of	O
what	O
is	O
filtered	O
out	O
,	O
Google	B-Application
takes	O
into	O
consideration	O
what	O
we	O
"	O
click	O
"	O
to	O
filter	O
out	O
what	O
comes	O
up	O
in	O
our	O
searches	O
.	O
</s>
<s>
In	O
addition	O
,	O
Facebook	B-Application
searches	O
are	O
not	O
necessarily	O
as	O
private	O
as	O
the	B-Application
Google	I-Application
ones	O
.	O
</s>
<s>
Facebook	B-Application
draws	O
on	O
the	O
more	O
public	O
self	O
and	O
users	O
share	O
what	O
other	O
people	O
want	O
to	O
see	O
.	O
</s>
<s>
Even	O
while	O
tagging	O
photographs	O
,	O
Facebook	B-Application
uses	O
personalization	O
and	O
face	O
recognition	O
that	O
will	O
automatically	O
assign	O
a	O
name	O
to	O
face	O
.	O
</s>
<s>
Facebook	B-Application
's	O
like	O
button	O
utilizes	O
its	O
users	O
to	O
do	O
their	O
own	O
personalization	O
for	O
the	O
website	O
.	O
</s>
<s>
What	O
posts	O
the	O
user	O
comments	O
on	O
or	O
likes	O
tells	O
Facebook	B-Application
what	O
type	O
of	O
posts	O
they	O
will	O
be	O
interested	O
in	O
for	O
the	O
future	O
.	O
</s>
<s>
In	O
addition	O
to	O
this	O
,	O
it	O
helps	O
them	O
predict	O
what	O
type	O
of	O
posts	O
they	O
will	O
“	O
comment	O
on	O
,	O
share	O
,	O
or	O
spam	O
in	O
the	O
future.	O
”	O
The	O
predictions	O
are	O
combined	O
to	O
produce	O
one	O
relevancy	O
score	O
which	O
helps	O
Facebook	B-Application
decide	O
what	O
to	O
show	O
you	O
and	O
what	O
to	O
filter	O
out	O
.	O
</s>
<s>
In	O
2016	O
,	O
Facebook	B-Application
introduced	O
reactions	O
(	O
Love	O
,	O
Thankful	O
,	O
Haha	O
,	O
Wow	O
,	O
Sad	O
,	O
and	O
Angry	O
)	O
in	O
addition	O
to	O
liking	O
a	O
post	O
.	O
</s>
<s>
“	O
Facebook	B-Application
has	O
learned	O
that	O
any	O
Reaction	O
left	O
on	O
a	O
post	O
is	O
a	O
strong	O
indicator	O
that	O
the	O
user	O
was	O
more	O
interested	O
in	O
that	O
post	O
than	O
any	O
other	O
‘	O
liked’	O
posts.	O
”	O
Facebook	B-Application
is	O
starting	O
to	O
weigh	O
reactions	O
the	O
same	O
way	O
as	O
likes	O
.	O
</s>
<s>
So	O
even	O
if	O
you	O
leave	O
the	O
“	O
angry	O
”	O
reaction	O
on	O
a	O
post	O
,	O
Facebook	B-Application
will	O
show	O
posts	O
on	O
the	O
user	O
's	O
feed	O
because	O
the	O
user	O
showed	O
an	O
interest	O
in	O
it	O
.	O
</s>
<s>
In	O
terms	O
of	O
Google	B-Application
,	O
users	O
are	O
provided	O
similar	O
websites	O
and	O
resources	O
based	O
on	O
what	O
they	O
initially	O
click	O
on	O
.	O
</s>
<s>
There	O
are	O
sites	O
like	O
Amazon	B-Application
and	O
personal	O
shopping	O
sites	O
also	O
use	O
other	O
peoples	O
history	O
in	O
order	O
to	O
serve	O
their	O
interests	O
better	O
.	O
</s>
<s>
Twitter	B-Application
also	O
uses	O
personalization	O
by	O
"	O
suggesting	O
"	O
other	O
people	O
to	O
follow	O
.	O
</s>
<s>
In	O
addition	O
,	O
based	O
on	O
who	O
one	O
"	O
follows	O
"	O
,	O
"	O
tweets	O
"	O
and	O
"	O
retweets	O
"	O
at	O
,	O
Twitter	B-Application
filters	O
out	O
suggestions	O
most	O
relevant	O
to	O
the	O
user	O
.	O
</s>
<s>
LinkedIn	B-Application
personalizes	O
search	B-Application
results	I-Application
at	O
two	O
levels	O
.	O
</s>
<s>
LinkedIn	B-Application
federated	B-Operating_System
search	I-Operating_System
exploits	O
user	O
intent	O
to	O
personalize	O
vertical	O
order	O
.	O
</s>
<s>
Mark	O
Zuckerberg	O
,	O
founder	O
of	O
Facebook	B-Application
,	O
believed	O
that	O
people	O
only	O
have	O
one	O
identity	O
.	O
</s>
<s>
Although	O
personalized	B-Application
search	I-Application
may	O
seem	O
helpful	O
,	O
it	O
is	O
not	O
a	O
very	O
accurate	O
representation	O
of	O
any	O
person	O
.	O
</s>
<s>
Search	B-Application
engines	I-Application
such	O
as	O
Google	B-Application
and	O
Yahoo	O
!	O
</s>
<s>
utilize	O
personalized	B-Application
search	I-Application
to	O
attract	O
possible	O
customers	O
to	O
products	O
that	O
fit	O
their	O
presumed	O
desires	O
.	O
</s>
<s>
Based	O
on	O
a	O
large	O
amount	O
of	O
collected	O
data	O
aggregated	O
from	O
an	O
individual	O
's	O
web	O
clicks	O
,	O
search	B-Application
engines	I-Application
can	O
use	O
personalized	B-Application
search	I-Application
to	O
put	O
advertisements	O
that	O
may	O
pique	O
the	O
interest	O
of	O
an	O
individual	O
.	O
</s>
<s>
Utilizing	O
personalized	B-Application
search	I-Application
can	O
help	O
consumers	O
find	O
what	O
they	O
want	O
faster	O
,	O
as	O
well	O
as	O
help	O
match	O
up	O
products	O
and	O
services	O
to	O
individuals	O
within	O
more	O
specialized	O
and/or	O
niche	O
markets	O
.	O
</s>
<s>
Using	O
personalized	B-Application
search	I-Application
allows	O
faster	O
product	O
and	O
service	O
discoveries	O
for	O
consumers	O
,	O
and	O
reduces	O
the	O
amount	O
of	O
necessary	O
advertisement	O
money	O
spent	O
to	O
reach	O
those	O
consumers	O
.	O
</s>
<s>
In	O
addition	O
,	O
utilizing	O
personalized	B-Application
search	I-Application
can	O
help	O
companies	O
determine	O
which	O
individuals	O
should	O
be	O
offered	O
online	O
coupon	O
codes	O
to	O
their	O
products	O
and/or	O
services	O
.	O
</s>
<s>
Aside	O
from	O
aiding	O
consumers	O
and	O
businesses	O
in	O
finding	O
one	O
another	O
,	O
the	O
search	B-Application
engines	I-Application
that	O
provide	O
personalized	B-Application
search	I-Application
benefit	O
greatly	O
.	O
</s>
<s>
In	O
turn	O
,	O
this	O
allows	O
search	B-Application
engines	I-Application
to	O
sell	O
more	O
advertisements	O
because	O
companies	O
understand	O
that	O
they	O
will	O
have	O
a	O
better	O
opportunity	O
to	O
sell	O
to	O
high	O
percentage	O
matched	O
individuals	O
then	O
medium	O
and	O
low	O
percentage	O
matched	O
individuals	O
.	O
</s>
<s>
This	O
aspect	O
of	O
personalized	B-Application
search	I-Application
angers	O
many	O
scholars	O
,	O
such	O
as	O
William	O
Badke	O
and	O
Eli	O
Pariser	O
,	O
because	O
they	O
believe	O
personalized	B-Application
search	I-Application
is	O
driven	O
by	O
the	O
desire	O
to	O
increase	O
advertisement	O
revenues	O
.	O
</s>
<s>
In	O
addition	O
,	O
they	O
believe	O
that	O
personalized	B-Application
search	I-Application
results	O
are	O
frequently	O
utilized	O
to	O
sway	O
individuals	O
into	O
using	O
products	O
and	O
services	O
that	O
are	O
offered	O
by	O
the	O
particular	O
search	B-Application
engine	I-Application
company	O
or	O
any	O
other	O
company	O
in	O
partnered	O
with	O
them	O
.	O
</s>
<s>
For	O
example	O
,	O
Google	B-Application
searching	O
any	O
company	O
with	O
at	O
least	O
one	O
brick-and-mortar	O
location	O
will	O
offer	O
a	O
map	O
portraying	O
the	O
closest	O
company	O
location	O
using	O
the	B-Application
Google	I-Application
Maps	O
service	O
as	O
the	O
first	O
result	O
to	O
the	O
query	O
.	O
</s>
<s>
Searching	O
the	O
word	O
"	O
shoes	O
"	O
using	O
the	B-Application
Google	I-Application
search	B-Application
engine	I-Application
will	O
offer	O
several	O
advertisements	O
to	O
shoe	O
companies	O
that	O
pay	O
Google	B-Application
to	O
link	O
their	O
website	O
as	O
a	O
first	O
result	O
to	O
consumer	O
's	O
queries	O
.	O
</s>
