►
Description
An example of taking the standard Hot Gym sample and using it to analyze stock prices over time.
http://nupic2015spring.challengepost.com/submissions/37840-ankur-saxena
A
Okay,
hey
everyone,
encore
and
first
I
want
to
say
thank
you
for
this
wonderful
experience
before
this
I've,
never
done
any
machine
learning
or
shown
your
network
or
anything
of
that
sort.
So
this
is
my
first
exposure
to
this
world
and
I
learned
so
much
and
I
learned
from
so
many
people
and
everyone
was
supportive,
so
I
want
to
say
thank
you
to
everyone
who
helped
me
answer,
questions
and
clear
out
the
concepts
behind
HTM
and
and
general
artificial
neural
networking
concepts.
A
So
what
I
did
here
was
I
had
a
time
series
data
on
a
stock,
ticker,
rlg
and
basically
I
used
Matt's,
Rajyam
implementation
and
fed
my
data
to
it
and
I
was
able
to
swarm
on
the
swarm
on
my
own
data
set
and
run
a
prediction
for
it.
So
here
it
is
working
and
the
data
sets
goes
back
from
1992
to
present
and
you
can
see
it
as
it's
predicting.
So
the
blue
is
the
actual
stock
price
and
green
is
that
predicted
and
it
seems
to
be
following
well
I
think
what
I
need
to
learn?
A
B
A
B
B
Trivial
sort
of
predictor
at
this
point
in
time,
if,
if
the
data
really
isn't
predictable,
that's
what
that's,
what
the
network
will
end
up
doing
it's
the
best
predictive
model
that
can
make
as
to
sort
of
say,
see
what
happened
recently
and
predict
that
and
most
people
think
back
stock
market
data.
It's
not
predictable!
So
but
it's
a
good
job.
He
did
a
lot.
You
accomplished
a
lot
and
one
when
we
command.