search

shachen / PLDS

The Penalized Linear Dynamical System Project
18 stars 12 forks source link

plot accuracy & time vs. iteration # #74

Open jovo opened 9 years ago

jovo commented 9 years ago

for accuracy, use accuracy of both E step and M step (don't count computing accuracy in the time calculation)

@yangyutu

jovo commented 9 years ago

how is this coming along? @vincentliuk @yangyutu

once this is done, let's chat again?

yangyutu commented 9 years ago

Hello, Josh,

I am getting confused. Are we supposed to make a new figure or update some previous figure?

Thanks.

On Tue, Aug 25, 2015 at 12:09 AM, joshua vogelstein < notifications@github.com> wrote:

how is this coming along? @vincentliuk https://github.com/vincentliuk @yangyutu https://github.com/yangyutu

once this is done, let's chat again?

— Reply to this email directly or view it on GitHub https://github.com/shachen/PLDS/issues/74#issuecomment-134468968.

jovo commented 9 years ago

new fig, 3 lines: parameter accuracy vs. iteration time vs. iteration prediction accuracy vs. iteration.

now that we understand time within iterations, we want to understand accuracy and time across iterations.

note: please include iteration 0, ie, the initialization phase.

is that clear?

On Tue, Aug 25, 2015 at 10:07 AM, Yuguang Yang notifications@github.com wrote:

Hello, Josh,

I am getting confused. Are we supposed to make a new figure or update some previous figure?

Thanks.

On Tue, Aug 25, 2015 at 12:09 AM, joshua vogelstein < notifications@github.com> wrote:

how is this coming along? @vincentliuk https://github.com/vincentliuk @yangyutu https://github.com/yangyutu

once this is done, let's chat again?

— Reply to this email directly or view it on GitHub https://github.com/shachen/PLDS/issues/74#issuecomment-134468968.

— Reply to this email directly or view it on GitHub https://github.com/shachen/PLDS/issues/74#issuecomment-134597141.

the glass is all full: half water, half air. openconnecto.me, jovo.me, office hours https://www.google.com/calendar/embed?src=vk2j3ugqcitvq223g7ivvq4hqo%40group.calendar.google.com&ctz=America/New_York

yangyutu commented 9 years ago

Hello, Josh, I have three questions: (1) time means computational time? (2) you said "use accuracy of both E step and M step", then for each iteration, I need to plot the accuracy for both E step and M step? Then there will be 5 lines rather than 3. (3) which data shall I use? I am planning to use the low dimensional setting as in Figure 1 for this plot.

jovo commented 9 years ago

yes. ok good start. thanks.

On Tue, Aug 25, 2015 at 11:39 AM, Yuguang Yang notifications@github.com wrote:

Hello, Josh, I have three questions: (1) time means computational time? (2) you said "use accuracy of both E step and M step", then for each iteration, I need to plot the accuracy for both E step and M step? Then there will be 5 lines rather than 3. (3) which data shall I use? I am planning to use the low dimensional setting as in Figure 1 for this plot.

— Reply to this email directly or view it on GitHub https://github.com/shachen/PLDS/issues/74#issuecomment-134633555.

the glass is all full: half water, half air. openconnecto.me, jovo.me, office hours https://www.google.com/calendar/embed?src=vk2j3ugqcitvq223g7ivvq4hqo%40group.calendar.google.com&ctz=America/New_York