Thanks Joel and Sören,
Joel, I've been reading up on Structure Learning off the pystruct project
documentation. It seems to describe my problem exactly. Thanks!
Dominic
On Sun, Jan 10, 2016 at 9:29 PM, Joel Nothman <joel.noth...@gmail.com>
wrote:
> I think you've misunderstood this one, Sören. This sounds like it is a
> structured learning problem, where the steps are the "target" of the
> learning task, and the result is the input example.
>
> Take, for instance, the natural language processing task of dependency
> parsing.
>
> The "result" of some latent syntactic process is "He quickly ate the cat",
> and the latent "steps" deriving that result are the dependency arcs:
>
> "He" is dependent to "ate"
> "quickly" is dependent to "ate"
> "the" is dependent to "cat"
> "cat" is dependent to "ate"
>
> Given a sentence, we want to recover such dependency arcs.
>
> This sort of structured learning problem is common in machine learning,
> but is generally beyond the scope of scikit-learn. (The related pystruct
> project by core scikit-learn dev Andreas Müller deals with this type of
> task.) One of the main challenges of structured learning problems is
> representing the target "steps" (i.e. the parse here) as, for example, some
> previously solved class of probabilistic graphical model, such as linear
> chain Conditional Random Fields, as well as representing features of the
> observed instance ("result") in a way that the machine learning model can
> learn to associate with the target.
>
> HTH
>
> Joel
>
> On 11 January 2016 at 01:38, Sören Wacker <soeren...@gmail.com> wrote:
>
>> Sounds like a classification problem. You can try to see the steps as
>> features and use classification methods e.g. decision tree to train a
>> model. But it depends on what "reconstruct the steps" and "result" means.
>>
>> Sören
>>
>>
>> On 01/09/2016 06:32 AM, Dominic Laflamme wrote:
>>
>> First I'd like to apologize in advanced for the "noobiness" nature of my
>> question...
>> I'd like to get some early guidance into which path I should take to help
>> me solve a problem using machine learning.
>>
>> The goal would be to use machine learning to "reconstruct the steps"
>> needed to take in order to get to a particular "result".
>> My assumption (perhaps misguided at this point) would be that I could
>> feed the system with a large amount of complete "examples" (steps +
>> result) to be eventually able to feed it a result in order to get a series
>> of steps.
>>
>> I know my question is broad, to say the least. But I feel that knowing
>> the right type of machine learning concept to apply to such a problem would
>> help me started.
>> Any insights on which category better suits my problem, and on whether my
>> assumptions are flawed?
>>
>> Any and all comments greatly appreciated.
>>
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Site24x7 APM Insight: Get Deep Visibility into Application Performance
>> APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
>> Monitor end-to-end web transactions and take corrective actions now
>> Troubleshoot faster and improve end-user experience. Signup
>> Now!http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140
>>
>>
>>
>> _______________________________________________
>> Scikit-learn-general mailing
>> listScikit-learn-general@lists.sourceforge.nethttps://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Site24x7 APM Insight: Get Deep Visibility into Application Performance
>> APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
>> Monitor end-to-end web transactions and take corrective actions now
>> Troubleshoot faster and improve end-user experience. Signup Now!
>> http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140
>> _______________________________________________
>> Scikit-learn-general mailing list
>> Scikit-learn-general@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>>
>>
>
>
> ------------------------------------------------------------------------------
> Site24x7 APM Insight: Get Deep Visibility into Application Performance
> APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> Monitor end-to-end web transactions and take corrective actions now
> Troubleshoot faster and improve end-user experience. Signup Now!
> http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140
> _______________________________________________
> Scikit-learn-general mailing list
> Scikit-learn-general@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
>
------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140
_______________________________________________
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
