How to find complete log likelihood for mixture of PPCA












1














In Appendix C of a paper by Michael E. Tipping and Christopher M. Bishop about mixture models for probabilistic PCA, the probability of a single data vector $mathbf{t}$ is expressed as a mixture of PCA models (equation 69):



$$
p(mathbf{t}) = sum_{i=1}^Mpi_i p(mathbf{t}|i)
$$



where $pi$ is the mixing proportion and $p(mathbf{t}|i)$ is a single probabilistic PCA model.



The model underlying the probabilistic PCA method is (equation 2)



$$
mathbf{t} = mathbf{Wx} + boldsymbolmu + boldsymbolepsilon.
$$
Where $mathbf{x}$ is a latent variable. By introducing a new set of variables $z_{ni}$ "labelling which model is responsible for generating each data point $mathbf{t}_n$", Bishop formulates the complete log likelihood as (equation 70):



$$
mathcal{L}_C = sum_{n=1}^Nsum_{i=1}^Mz_{ni}ln{pi_ip(mathbf{t}_n, mathbf{x}_{ni})}.
$$
I would like to understand how he derives this expression as he doesn't provide a solution himself. How is this expression for the complete log likelihood found?






probability-theory machine-learning







share|cite|improve this question



























    1














    In Appendix C of a paper by Michael E. Tipping and Christopher M. Bishop about mixture models for probabilistic PCA, the probability of a single data vector $mathbf{t}$ is expressed as a mixture of PCA models (equation 69):



    $$
    p(mathbf{t}) = sum_{i=1}^Mpi_i p(mathbf{t}|i)
    $$



    where $pi$ is the mixing proportion and $p(mathbf{t}|i)$ is a single probabilistic PCA model.



    The model underlying the probabilistic PCA method is (equation 2)



    $$
    mathbf{t} = mathbf{Wx} + boldsymbolmu + boldsymbolepsilon.
    $$
    Where $mathbf{x}$ is a latent variable. By introducing a new set of variables $z_{ni}$ "labelling which model is responsible for generating each data point $mathbf{t}_n$", Bishop formulates the complete log likelihood as (equation 70):



    $$
    mathcal{L}_C = sum_{n=1}^Nsum_{i=1}^Mz_{ni}ln{pi_ip(mathbf{t}_n, mathbf{x}_{ni})}.
    $$
    I would like to understand how he derives this expression as he doesn't provide a solution himself. How is this expression for the complete log likelihood found?






    probability-theory machine-learning







    share|cite|improve this question

























      1












      1








      1







      In Appendix C of a paper by Michael E. Tipping and Christopher M. Bishop about mixture models for probabilistic PCA, the probability of a single data vector $mathbf{t}$ is expressed as a mixture of PCA models (equation 69):



      $$
      p(mathbf{t}) = sum_{i=1}^Mpi_i p(mathbf{t}|i)
      $$



      where $pi$ is the mixing proportion and $p(mathbf{t}|i)$ is a single probabilistic PCA model.



      The model underlying the probabilistic PCA method is (equation 2)



      $$
      mathbf{t} = mathbf{Wx} + boldsymbolmu + boldsymbolepsilon.
      $$
      Where $mathbf{x}$ is a latent variable. By introducing a new set of variables $z_{ni}$ "labelling which model is responsible for generating each data point $mathbf{t}_n$", Bishop formulates the complete log likelihood as (equation 70):



      $$
      mathcal{L}_C = sum_{n=1}^Nsum_{i=1}^Mz_{ni}ln{pi_ip(mathbf{t}_n, mathbf{x}_{ni})}.
      $$
      I would like to understand how he derives this expression as he doesn't provide a solution himself. How is this expression for the complete log likelihood found?






      probability-theory machine-learning







      share|cite|improve this question













      In Appendix C of a paper by Michael E. Tipping and Christopher M. Bishop about mixture models for probabilistic PCA, the probability of a single data vector $mathbf{t}$ is expressed as a mixture of PCA models (equation 69):



      $$
      p(mathbf{t}) = sum_{i=1}^Mpi_i p(mathbf{t}|i)
      $$



      where $pi$ is the mixing proportion and $p(mathbf{t}|i)$ is a single probabilistic PCA model.



      The model underlying the probabilistic PCA method is (equation 2)



      $$
      mathbf{t} = mathbf{Wx} + boldsymbolmu + boldsymbolepsilon.
      $$
      Where $mathbf{x}$ is a latent variable. By introducing a new set of variables $z_{ni}$ "labelling which model is responsible for generating each data point $mathbf{t}_n$", Bishop formulates the complete log likelihood as (equation 70):



      $$
      mathcal{L}_C = sum_{n=1}^Nsum_{i=1}^Mz_{ni}ln{pi_ip(mathbf{t}_n, mathbf{x}_{ni})}.
      $$
      I would like to understand how he derives this expression as he doesn't provide a solution himself. How is this expression for the complete log likelihood found?






      probability-theory machine-learning




      probability-theory machine-learning






      share|cite|improve this question













      share|cite|improve this question











      share|cite|improve this question




      share|cite|improve this question










      asked Jan 4 at 11:07









      SandiSandi

      255112




      255112






















          1 Answer
          1






          active

          oldest

          votes


















          0














          Let's concentrate for the time being on the $n^{rm th}$ datapoint, $mathbf t_n$. Suppose this datapoint is generated from the $i_n^{rm th}$ model. Then $$ z_{ni} = begin{cases} 1 & {rm if } i = i_n \ 0 & {rm otherwise}end{cases}.$$Thus we have
          $$sum_{i=1}^M z_{ni} ln left(pi_i p(mathbf t_n ,mathbf x_{ni}) right) = ln left( pi_{i_n} p(mathbf t_n , mathbf x_{n{i_n}})right).$$
          The expression on the right-hand side is log-likelihood for the $n$th datapoint. To spell it out:





          • $pi_{i_n}$ is the probability that the $n^{rm th}$ datapoint is generated by the $i_n^{rm th}$ model.


          • $p(mathbf t_n, mathbf x_{ni_n})$ is the probability of encountering this particular latent vector $mathbf x_{ni_n}$ and this particular visible vector $mathbf t_n$ for the $n^{rm th}$ datapoint, given that this datapoint is generated from the the $i_n^{rm th}$ model. [In fact, $$p(mathbf t_n, mathbf x_{ni_n}) = mathcal N(mathbf x_{ni_n} | mathbf 0, mathbf I) times mathcal N(mathbf t_n - mathbf W mathbf x_{ni_n} - mathbf mu_n |mathbf 0, sigma_n^2 mathbf I),$$ assuming that $mathbf epsilon_{n} sim mathcal N(mathbf 0, sigma_n^2 mathbf I)$.]


          Since the datapoints are generated independently, the log-likelihood for the entire dataset is a sum over the log-likelihoods for the individual datapoints, giving the desired result.






          share|cite|improve this answer





















            Your Answer





            StackExchange.ifUsing("editor", function () {
            return StackExchange.using("mathjaxEditing", function () {
            StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix) {
            StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
            });
            });
            }, "mathjax-editing");

            StackExchange.ready(function() {
            var channelOptions = {
            tags: "".split(" "),
            id: "69"
            };
            initTagRenderer("".split(" "), "".split(" "), channelOptions);

            StackExchange.using("externalEditor", function() {
            // Have to fire editor after snippets, if snippets enabled
            if (StackExchange.settings.snippets.snippetsEnabled) {
            StackExchange.using("snippets", function() {
            createEditor();
            });
            }
            else {
            createEditor();
            }
            });

            function createEditor() {
            StackExchange.prepareEditor({
            heartbeatType: 'answer',
            autoActivateHeartbeat: false,
            convertImagesToLinks: true,
            noModals: true,
            showLowRepImageUploadWarning: true,
            reputationToPostImages: 10,
            bindNavPrevention: true,
            postfix: "",
            imageUploader: {
            brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
            contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
            allowUrls: true
            },
            noCode: true, onDemand: true,
            discardSelector: ".discard-answer"
            ,immediatelyShowMarkdownHelp:true
            });


            }
            });














            draft saved

            draft discarded


















            StackExchange.ready(
            function () {
            StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3061531%2fhow-to-find-complete-log-likelihood-for-mixture-of-ppca%23new-answer', 'question_page');
            }
            );

            Post as a guest















            Required, but never shown

























            1 Answer
            1






            active

            oldest

            votes








            1 Answer
            1






            active

            oldest

            votes









            active

            oldest

            votes






            active

            oldest

            votes









            0














            Let's concentrate for the time being on the $n^{rm th}$ datapoint, $mathbf t_n$. Suppose this datapoint is generated from the $i_n^{rm th}$ model. Then $$ z_{ni} = begin{cases} 1 & {rm if } i = i_n \ 0 & {rm otherwise}end{cases}.$$Thus we have
            $$sum_{i=1}^M z_{ni} ln left(pi_i p(mathbf t_n ,mathbf x_{ni}) right) = ln left( pi_{i_n} p(mathbf t_n , mathbf x_{n{i_n}})right).$$
            The expression on the right-hand side is log-likelihood for the $n$th datapoint. To spell it out:





            • $pi_{i_n}$ is the probability that the $n^{rm th}$ datapoint is generated by the $i_n^{rm th}$ model.


            • $p(mathbf t_n, mathbf x_{ni_n})$ is the probability of encountering this particular latent vector $mathbf x_{ni_n}$ and this particular visible vector $mathbf t_n$ for the $n^{rm th}$ datapoint, given that this datapoint is generated from the the $i_n^{rm th}$ model. [In fact, $$p(mathbf t_n, mathbf x_{ni_n}) = mathcal N(mathbf x_{ni_n} | mathbf 0, mathbf I) times mathcal N(mathbf t_n - mathbf W mathbf x_{ni_n} - mathbf mu_n |mathbf 0, sigma_n^2 mathbf I),$$ assuming that $mathbf epsilon_{n} sim mathcal N(mathbf 0, sigma_n^2 mathbf I)$.]


            Since the datapoints are generated independently, the log-likelihood for the entire dataset is a sum over the log-likelihoods for the individual datapoints, giving the desired result.






            share|cite|improve this answer


























              0














              Let's concentrate for the time being on the $n^{rm th}$ datapoint, $mathbf t_n$. Suppose this datapoint is generated from the $i_n^{rm th}$ model. Then $$ z_{ni} = begin{cases} 1 & {rm if } i = i_n \ 0 & {rm otherwise}end{cases}.$$Thus we have
              $$sum_{i=1}^M z_{ni} ln left(pi_i p(mathbf t_n ,mathbf x_{ni}) right) = ln left( pi_{i_n} p(mathbf t_n , mathbf x_{n{i_n}})right).$$
              The expression on the right-hand side is log-likelihood for the $n$th datapoint. To spell it out:





              • $pi_{i_n}$ is the probability that the $n^{rm th}$ datapoint is generated by the $i_n^{rm th}$ model.


              • $p(mathbf t_n, mathbf x_{ni_n})$ is the probability of encountering this particular latent vector $mathbf x_{ni_n}$ and this particular visible vector $mathbf t_n$ for the $n^{rm th}$ datapoint, given that this datapoint is generated from the the $i_n^{rm th}$ model. [In fact, $$p(mathbf t_n, mathbf x_{ni_n}) = mathcal N(mathbf x_{ni_n} | mathbf 0, mathbf I) times mathcal N(mathbf t_n - mathbf W mathbf x_{ni_n} - mathbf mu_n |mathbf 0, sigma_n^2 mathbf I),$$ assuming that $mathbf epsilon_{n} sim mathcal N(mathbf 0, sigma_n^2 mathbf I)$.]


              Since the datapoints are generated independently, the log-likelihood for the entire dataset is a sum over the log-likelihoods for the individual datapoints, giving the desired result.






              share|cite|improve this answer
























                0












                0








                0






                Let's concentrate for the time being on the $n^{rm th}$ datapoint, $mathbf t_n$. Suppose this datapoint is generated from the $i_n^{rm th}$ model. Then $$ z_{ni} = begin{cases} 1 & {rm if } i = i_n \ 0 & {rm otherwise}end{cases}.$$Thus we have
                $$sum_{i=1}^M z_{ni} ln left(pi_i p(mathbf t_n ,mathbf x_{ni}) right) = ln left( pi_{i_n} p(mathbf t_n , mathbf x_{n{i_n}})right).$$
                The expression on the right-hand side is log-likelihood for the $n$th datapoint. To spell it out:





                • $pi_{i_n}$ is the probability that the $n^{rm th}$ datapoint is generated by the $i_n^{rm th}$ model.


                • $p(mathbf t_n, mathbf x_{ni_n})$ is the probability of encountering this particular latent vector $mathbf x_{ni_n}$ and this particular visible vector $mathbf t_n$ for the $n^{rm th}$ datapoint, given that this datapoint is generated from the the $i_n^{rm th}$ model. [In fact, $$p(mathbf t_n, mathbf x_{ni_n}) = mathcal N(mathbf x_{ni_n} | mathbf 0, mathbf I) times mathcal N(mathbf t_n - mathbf W mathbf x_{ni_n} - mathbf mu_n |mathbf 0, sigma_n^2 mathbf I),$$ assuming that $mathbf epsilon_{n} sim mathcal N(mathbf 0, sigma_n^2 mathbf I)$.]


                Since the datapoints are generated independently, the log-likelihood for the entire dataset is a sum over the log-likelihoods for the individual datapoints, giving the desired result.






                share|cite|improve this answer












                Let's concentrate for the time being on the $n^{rm th}$ datapoint, $mathbf t_n$. Suppose this datapoint is generated from the $i_n^{rm th}$ model. Then $$ z_{ni} = begin{cases} 1 & {rm if } i = i_n \ 0 & {rm otherwise}end{cases}.$$Thus we have
                $$sum_{i=1}^M z_{ni} ln left(pi_i p(mathbf t_n ,mathbf x_{ni}) right) = ln left( pi_{i_n} p(mathbf t_n , mathbf x_{n{i_n}})right).$$
                The expression on the right-hand side is log-likelihood for the $n$th datapoint. To spell it out:





                • $pi_{i_n}$ is the probability that the $n^{rm th}$ datapoint is generated by the $i_n^{rm th}$ model.


                • $p(mathbf t_n, mathbf x_{ni_n})$ is the probability of encountering this particular latent vector $mathbf x_{ni_n}$ and this particular visible vector $mathbf t_n$ for the $n^{rm th}$ datapoint, given that this datapoint is generated from the the $i_n^{rm th}$ model. [In fact, $$p(mathbf t_n, mathbf x_{ni_n}) = mathcal N(mathbf x_{ni_n} | mathbf 0, mathbf I) times mathcal N(mathbf t_n - mathbf W mathbf x_{ni_n} - mathbf mu_n |mathbf 0, sigma_n^2 mathbf I),$$ assuming that $mathbf epsilon_{n} sim mathcal N(mathbf 0, sigma_n^2 mathbf I)$.]


                Since the datapoints are generated independently, the log-likelihood for the entire dataset is a sum over the log-likelihoods for the individual datapoints, giving the desired result.







                share|cite|improve this answer












                share|cite|improve this answer



                share|cite|improve this answer










                answered yesterday









                Kenny WongKenny Wong

                18.3k21438




                18.3k21438






























                    draft saved

                    draft discarded




















































                    Thanks for contributing an answer to Mathematics Stack Exchange!


                    • Please be sure to answer the question. Provide details and share your research!

                    But avoid …



                    • Asking for help, clarification, or responding to other answers.

                    • Making statements based on opinion; back them up with references or personal experience.


                    Use MathJax to format equations. MathJax reference.


                    To learn more, see our tips on writing great answers.





                    Some of your past answers have not been well-received, and you're in danger of being blocked from answering.


                    Please pay close attention to the following guidance:


                    • Please be sure to answer the question. Provide details and share your research!

                    But avoid …



                    • Asking for help, clarification, or responding to other answers.

                    • Making statements based on opinion; back them up with references or personal experience.


                    To learn more, see our tips on writing great answers.




                    draft saved


                    draft discarded














                    StackExchange.ready(
                    function () {
                    StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3061531%2fhow-to-find-complete-log-likelihood-for-mixture-of-ppca%23new-answer', 'question_page');
                    }
                    );

                    Post as a guest















                    Required, but never shown





















































                    Required, but never shown














                    Required, but never shown












                    Required, but never shown







                    Required, but never shown

































                    Required, but never shown














                    Required, but never shown












                    Required, but never shown







                    Required, but never shown







                    -

                    Popular posts from this blog

                    An IMO inspired problem

                    Image
                    14 4 This problem from IMO 1988 is said to be one of the most elegant ones in functional equations. Problem : The function $f$ is defined on the set of all positive integers as follows: begin{align} f(1) = 1, f(3) &= 3, f(2n) = f(n), \ f(4n+1) &= 2f(2n+1) - f(n), \ f(4n+3) &= 3 f(2n+1) - 2f(n) end{align} Find the number of $n$ with $f(n) = n, 1 leq n leq 1988$. The main idea towards the solution is realizing that these conditions stand for the fact that $f(n)$ just reverses the digits in the binary representation of the number $n$. So, essentially the solution is finding the number of binary pallindromes $leq 1988_{10}$. My question is the following : How to reformulate the problem so that $f(n)$ reverses the digits of $n$ in its ternary representation. Or even better, can we reformulate it for a
                    Read more

                    Management

                    Image
                    Management From Wikipedia, the free encyclopedia Jump to navigation Jump to search "Manager" redirects here. For other uses, see Management (disambiguation) and Manager (disambiguation). An organization chart for the United States Coast Guard shows the hierarchy of managerial roles in that organization. Business administration Management of a business Accounting Management accounting Financial accounting Financial audit Business entities Cooperative Corporation Limited liability company Partnership Sole proprietorship State-owned enterprise Corporate governance Annual general meeting Board of directors Supervisory board Advisory board Audit committee Corporate law Commercial law Constitutional documents Contract Corporate crime Corporate liability Insolvency law International trade law Mergers and acquisitions Economics Commodity Pu
                    Read more

                    Has there ever been an instance of an active nuclear power plant within or near a war zone?

                    Image
                    36 3 The question Has there ever been a military combat going on around an active (running or in outage but not decommissioned) commercial nuclear power plant? Or dangerously near it? If yes, how did the plant and its personnel fare in such situation? Why I ask Let's assume for the scope of this question modern commercial nuclear power plants (CANDU, PWR, BWR of generation II plus all of gen III and newer) are usually reasonably safe to operate given that the engineering has been done right and there is a whole cohort of very well trained personnel on-site at all times, well rested and with considerable resources on their hands. Things hard to get with an armed conflict like the recent Syrian war raging around. I can't seem to find any reference or a comprehensive article about this, hence this question.
                    Read more