Note: I'm a stats novice, please let me know if any of these terms are unclear or misused and I'll update the question!

I'd like to predict future values of a time series. More precisely, I'm interested in the distribution of the next value.

I've chosen a normal/Gaussian prior. I see that I can calculate my new posterior distribution by solving the conjugate prior (https://en.wikipedia.org/wiki/Conjugate_prior, Continuous Distributions section, normal with known variance).

This seems to treat all observations of my time series equally. What if my time series has an underlying trend upward or down - is there a way to represent and quantify this?

For example, I've taken an existing time series with ~30 data points in the range of 30-50. I now add 10 fake data points that are double the max observed value. This barely moves my distribution's mean by ~4. This is across a range of mean/variance combinations for my prior. With 25% of my observations so far above the prior's mean, and especially because they all occur recently, my intuition is that I should see a large shift in posterior mean and posterior variance.

Is my intuition wrong, or is there something like a "Bayes moving normal" that I can use?

  • $\begingroup$ Your issue has less to do with Bayesian estimation and more to do with your choice of time series model. If you believe your model has a trend you can model your data as a function of time or use an auto - regressive model with a drift. Exponential weighting can also be used in most settings. I would encourage you to look into ARIMA models as they may accomplish a lot of what you are looking for. These models can be estimated in either a Bayesian or frequentest setting. $\endgroup$ – Zachary Blumenfeld Jan 22 '16 at 2:39
  • $\begingroup$ I agree with Zachary. As soon as I read 'I've chosen a normal/gaussian prior' I asked myself: "prior for what? What's the model?" Also, in the paragraph after that, well, weighting observations differently isn't really connected to the idea of a trend explaining your time series. Also there are different types of trend (stochastic and deterministic). Hope this helps. $\endgroup$ – Taylor Jan 22 '16 at 4:34

Your question about Bayesian Time Series falls perhaps into Bayesian state-space models (Bayesian Filtering) and the problem could be formulated under the same structure.

Within Bayesian state-space models, under the assumption of linearity, you could structure you dynamics as in a Kalman Filter (the "easiest" you can get).

$$ x_{k+1} = Ax + w_k \\ y_{k+1} = Hx + v_k $$

where, if you assume the latent variables are the weights given to the AR factors (down until $t-n$):

$$x_k = x_{k-1} = x = \begin{bmatrix} a_0 \\ \vdots \\ a_n \end{bmatrix} = a$$


$$H = [y_{k} \quad y_{k-1} \quad .. ] \dot \quad a$$

A Bayesian state-space model gives you:

  • the possibility to fine-tune the weight given to the latest measurements
  • the possibility to have the weights on the AR factors adapted in response to new data

Your questions seemed concerned about fine-tuning how much weight should be given to the latest measurements. If we model the measurement noise $v_k$ where $v_k \sim N(0, e_v)$ "smaller" (*) than the model uncertainty, the filter will rely on the latest measurement and update the AR factors $a$ faster in response to high prediction errors

You can see how some AR, MA and ARMA models can be refactored into state-space models here

(*) lot's to say about this - I am trying to convey the idea here. See "Kalman gain tuning"

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.