tag:blogger.com,1999:blog-3584326757943285743.post8657111687353474855..comments2017-04-06T00:10:09.838-07:00Comments on Research in Reading: Pushing the boundaries of what we knowPhilip Maybanknoreply@blogger.comBlogger3125tag:blogger.com,1999:blog-3584326757943285743.post-56083881128122016262017-04-06T00:10:09.838-07:002017-04-06T00:10:09.838-07:00Thank you very much for the references, now I thin...Thank you very much for the references, now I think I have a better idea of where to start studying!valentina carapellahttps://www.blogger.com/profile/16268584042098552893noreply@blogger.comtag:blogger.com,1999:blog-3584326757943285743.post-4427724826868241812017-03-30T07:25:03.441-07:002017-03-30T07:25:03.441-07:00Thanks for the question! One specific issue that ...Thanks for the question! One specific issue that can come up when you are deciding what variables to include in your model is that more complex models tend to fit data getter than simpler ones. Sometimes this is because they more accurately describe the underlying physical process / signal, but sometimes they give a better fit because they are fitting the noise better.<br /><br />There are two different paths you can go down to address this issue.<br /><br />(1) Validate the model by calculating out-of-sample prediction errors.<br />(2) Calculate / estimate marginal likelihoods and use Bayes factors to compare models.<br /><br />Both of these approaches would penalize an overly complex model.<br /><br />In terms of references, the following books may be useful,<br /><br />Hastie, Trevor, Tibshirani, Robert, Friedman, J. (2009). The Elements of Statistical Learning (2nd ed.). Springer<br /><br />Gelman, A., et al. (2014). Bayesian Data Analysis (3rd ed.). CRC Press.<br /><br />The following paper is a good example of the marginal likelihood approach applied to differential equation models. It uses fairly recently developed methods.<br /><br />Penny, W., & Sengupta, B. (2016). Annealed Importance Sampling for Neural Mass Models. PLoS Comput BiolPhilip Maybankhttps://www.blogger.com/profile/00240300518426261149noreply@blogger.comtag:blogger.com,1999:blog-3584326757943285743.post-72996844954061304492017-03-20T02:35:07.706-07:002017-03-20T02:35:07.706-07:00This is a really interesting reflection, thanks fo...This is a really interesting reflection, thanks for sharing it. I am particularly interested in the point you make about models that are not really improved by including all the current knowledge available. It is challenging to understand what variables really have a weight on the mechanisms one is trying to capture, and what can be discarded in relation to the specific research question. In my limited experience I employ sensitivity analysis, but this is only a limited answer. How do you proceed when selecting or discarding variables? Can you recommend further reads on this topic?valentina carapellahttps://www.blogger.com/profile/16268584042098552893noreply@blogger.com