Analytical value of multivariate normal posterior
$begingroup$
Suppose I have the following Bayesian Network:
It's given by the following relations:
$$begin{aligned}X_1&sim mathcal N(mu, 1/sigma^2)\
forall k, 2leq kleq n: X_k|X_{k-1}&sim mathcal N(x_{k-1}, 1/lambda^2)\
forall i, 1leq ileq n: Z_i|X_i&simmathcal N(x_{i}, text I_i/delta^2)
end{aligned}$$
It is possible to find the expected value of $boldsymbol X|boldsymbol Z$ by minimising the following error function:
$$sum_{i=1}^ntext I_i(x_i-z_i)^2 + frac{delta^2}{lambda^2}sum_{i=2}^n(x_i-x_{i-1})^2 + frac{delta^2}{sigma^2}(x_1-mu)^2$$
(You can find the proof here.)
The solutions will necessarily follow these equalities:
$$begin{aligned}
hat x_1 &= frac{frac{delta^2}{lambda^2}hat x_2 + text I_1z_1 + frac{delta^2}{sigma^2}mu}{frac{delta^2}{lambda^2} + text I_1 + frac{delta^2}{sigma^2}} \
hat x_n &= frac{text I_nz_n + frac{delta^2}{sigma^2}x_{n-1}}{text I_n + frac{delta^2}{sigma^2}} \
forall i in {2, 3, ..., n-1}: hat x_i &= frac{frac{delta^2}{lambda^2}(hat x_{i + 1} + hat x_{i - 1}) + text I_iz_i}{2frac{delta^2}{lambda^2} + text I_i }
end{aligned}$$
It's possible to find $hat x_n$ by replacing $hat x_1$ in $hat x_2$ (and then $hat x_2$ will only depend on $hat x_3$), then replacing $hat x_2$ in $hat x_3$, and so on, until we reach $hat x_n$, at which point it's numbers all the way down and we have an actual value.
Is there an easier way to solve this, though? Or an implementation of this algorithm in R or something? I've been trying to think of an R implementation of this that doesn't consist of fully solving the optimisation problem and then taking the last component of the resulting vector, since that takes longer than I'd wish, and given that there's an "analytical" solution to the problem I'd hope to be able to figure out how to implement it.
probability-theory optimization kalman-filter
asked Dec 11 '18 at 20:33
Pedro CarvalhoPedro Carvalho
399215
$endgroup$
add a comment |
$begingroup$
Suppose I have the following Bayesian Network:
It's given by the following relations:
$$begin{aligned}X_1&sim mathcal N(mu, 1/sigma^2)\
forall k, 2leq kleq n: X_k|X_{k-1}&sim mathcal N(x_{k-1}, 1/lambda^2)\
forall i, 1leq ileq n: Z_i|X_i&simmathcal N(x_{i}, text I_i/delta^2)
end{aligned}$$
It is possible to find the expected value of $boldsymbol X|boldsymbol Z$ by minimising the following error function:
$$sum_{i=1}^ntext I_i(x_i-z_i)^2 + frac{delta^2}{lambda^2}sum_{i=2}^n(x_i-x_{i-1})^2 + frac{delta^2}{sigma^2}(x_1-mu)^2$$
(You can find the proof here.)
The solutions will necessarily follow these equalities:
$$begin{aligned}
hat x_1 &= frac{frac{delta^2}{lambda^2}hat x_2 + text I_1z_1 + frac{delta^2}{sigma^2}mu}{frac{delta^2}{lambda^2} + text I_1 + frac{delta^2}{sigma^2}} \
hat x_n &= frac{text I_nz_n + frac{delta^2}{sigma^2}x_{n-1}}{text I_n + frac{delta^2}{sigma^2}} \
forall i in {2, 3, ..., n-1}: hat x_i &= frac{frac{delta^2}{lambda^2}(hat x_{i + 1} + hat x_{i - 1}) + text I_iz_i}{2frac{delta^2}{lambda^2} + text I_i }
end{aligned}$$
It's possible to find $hat x_n$ by replacing $hat x_1$ in $hat x_2$ (and then $hat x_2$ will only depend on $hat x_3$), then replacing $hat x_2$ in $hat x_3$, and so on, until we reach $hat x_n$, at which point it's numbers all the way down and we have an actual value.
Is there an easier way to solve this, though? Or an implementation of this algorithm in R or something? I've been trying to think of an R implementation of this that doesn't consist of fully solving the optimisation problem and then taking the last component of the resulting vector, since that takes longer than I'd wish, and given that there's an "analytical" solution to the problem I'd hope to be able to figure out how to implement it.
probability-theory optimization kalman-filter
asked Dec 11 '18 at 20:33
Pedro CarvalhoPedro Carvalho
399215
$endgroup$
add a comment |
$begingroup$
Suppose I have the following Bayesian Network:
It's given by the following relations:
$$begin{aligned}X_1&sim mathcal N(mu, 1/sigma^2)\
forall k, 2leq kleq n: X_k|X_{k-1}&sim mathcal N(x_{k-1}, 1/lambda^2)\
forall i, 1leq ileq n: Z_i|X_i&simmathcal N(x_{i}, text I_i/delta^2)
end{aligned}$$
It is possible to find the expected value of $boldsymbol X|boldsymbol Z$ by minimising the following error function:
$$sum_{i=1}^ntext I_i(x_i-z_i)^2 + frac{delta^2}{lambda^2}sum_{i=2}^n(x_i-x_{i-1})^2 + frac{delta^2}{sigma^2}(x_1-mu)^2$$
(You can find the proof here.)
The solutions will necessarily follow these equalities:
$$begin{aligned}
hat x_1 &= frac{frac{delta^2}{lambda^2}hat x_2 + text I_1z_1 + frac{delta^2}{sigma^2}mu}{frac{delta^2}{lambda^2} + text I_1 + frac{delta^2}{sigma^2}} \
hat x_n &= frac{text I_nz_n + frac{delta^2}{sigma^2}x_{n-1}}{text I_n + frac{delta^2}{sigma^2}} \
forall i in {2, 3, ..., n-1}: hat x_i &= frac{frac{delta^2}{lambda^2}(hat x_{i + 1} + hat x_{i - 1}) + text I_iz_i}{2frac{delta^2}{lambda^2} + text I_i }
end{aligned}$$
It's possible to find $hat x_n$ by replacing $hat x_1$ in $hat x_2$ (and then $hat x_2$ will only depend on $hat x_3$), then replacing $hat x_2$ in $hat x_3$, and so on, until we reach $hat x_n$, at which point it's numbers all the way down and we have an actual value.
Is there an easier way to solve this, though? Or an implementation of this algorithm in R or something? I've been trying to think of an R implementation of this that doesn't consist of fully solving the optimisation problem and then taking the last component of the resulting vector, since that takes longer than I'd wish, and given that there's an "analytical" solution to the problem I'd hope to be able to figure out how to implement it.
probability-theory optimization kalman-filter
asked Dec 11 '18 at 20:33
Pedro CarvalhoPedro Carvalho
399215
$endgroup$
Suppose I have the following Bayesian Network:
It's given by the following relations:
$$begin{aligned}X_1&sim mathcal N(mu, 1/sigma^2)\
forall k, 2leq kleq n: X_k|X_{k-1}&sim mathcal N(x_{k-1}, 1/lambda^2)\
forall i, 1leq ileq n: Z_i|X_i&simmathcal N(x_{i}, text I_i/delta^2)
end{aligned}$$
It is possible to find the expected value of $boldsymbol X|boldsymbol Z$ by minimising the following error function:
$$sum_{i=1}^ntext I_i(x_i-z_i)^2 + frac{delta^2}{lambda^2}sum_{i=2}^n(x_i-x_{i-1})^2 + frac{delta^2}{sigma^2}(x_1-mu)^2$$
(You can find the proof here.)
The solutions will necessarily follow these equalities:
$$begin{aligned}
hat x_1 &= frac{frac{delta^2}{lambda^2}hat x_2 + text I_1z_1 + frac{delta^2}{sigma^2}mu}{frac{delta^2}{lambda^2} + text I_1 + frac{delta^2}{sigma^2}} \
hat x_n &= frac{text I_nz_n + frac{delta^2}{sigma^2}x_{n-1}}{text I_n + frac{delta^2}{sigma^2}} \
forall i in {2, 3, ..., n-1}: hat x_i &= frac{frac{delta^2}{lambda^2}(hat x_{i + 1} + hat x_{i - 1}) + text I_iz_i}{2frac{delta^2}{lambda^2} + text I_i }
end{aligned}$$
It's possible to find $hat x_n$ by replacing $hat x_1$ in $hat x_2$ (and then $hat x_2$ will only depend on $hat x_3$), then replacing $hat x_2$ in $hat x_3$, and so on, until we reach $hat x_n$, at which point it's numbers all the way down and we have an actual value.
Is there an easier way to solve this, though? Or an implementation of this algorithm in R or something? I've been trying to think of an R implementation of this that doesn't consist of fully solving the optimisation problem and then taking the last component of the resulting vector, since that takes longer than I'd wish, and given that there's an "analytical" solution to the problem I'd hope to be able to figure out how to implement it.
probability-theory optimization kalman-filter
probability-theory optimization kalman-filter
asked Dec 11 '18 at 20:33
Pedro CarvalhoPedro Carvalho
399215
asked Dec 11 '18 at 20:33
Pedro CarvalhoPedro Carvalho
399215
asked Dec 11 '18 at 20:33
Pedro CarvalhoPedro Carvalho
399215
asked Dec 11 '18 at 20:33
Pedro CarvalhoPedro Carvalho
399215
asked Dec 11 '18 at 20:33
Pedro CarvalhoPedro Carvalho
399215
399215
add a comment |
add a comment |
0
active
oldest
votes
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
});
}
});
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3035779%2fanalytical-value-of-multivariate-normal-posterior%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
0
active
oldest
votes
0
active
oldest
votes
active
oldest
votes
active
oldest
votes
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.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3035779%2fanalytical-value-of-multivariate-normal-posterior%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
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
