New answers tagged

0

There is a whole discussion about using bootstrap in time series data. In dynamic models like GARCH you associate today's volatility with yesterday's and so on. When you sample with replacement the ordering collapses and you usually end up with data having roughly the same moments with the original but different dynamic characteristics. In time series ...


0

In Random subspace method, we subsample the features, and base learners can be anything. In random forests, we subsample the training set, the features and use decision trees as base learners. All of these subsampling are with replacement in both methods.


8

@gunes has already given the answer, that random forests do work well with interactions. It may help to give a practical example, though. For that, we will need some toy data set.seed(2021) n = 5000 x1 <- rnorm(2*n) x2 <- runif(2*n) interaction <- x1 * x2 pseudointeraction <- sample(interaction) x3 <- rbeta(2*n, 2, 10) epsilon <- rt(2*n, df ...


7

Yes, tree based methods are good at detecting interactions, but not always. For example, using $x_1\lessgtr0$ and $x_2\lessgtr0$ in subsequent levels of the tree would be equivalent to using $x_1x_2\lessgtr0$ on one level. That said, since you set a max depth hyperparameter in random forests, adding promising interactions will decrease your overall depth and ...


1

I don't know if there is a standard way, but here is a paper that discusses one approach. (DISCLAIMER: I know the authors, so take it as you will). There is an R package implementing the test and a description and examples of how to apply it. That paper also cites and compares with other approaches such as edge cross-validation and network cross-validation (...


1

You are right, modularity maximization is very unstable: just renumbering vertices, or removing/adding one edge may have a huge impact on obtained results. This was shown for instance in: Static community detection algorithms for evolving networks However, iterating community detection leads to very stable pieces in most practical cases: Stable community ...


1

To begin, I confess I'm not familiar with either the SAS/binomial method or (just from your bit of R code) the exact type of bootstrap CI you are using. I notice from your plot that the median is not near the middle of either of these types of CIs. There are various correct methods for making CIs, based on different assumptions and criteria, so both kinds of ...


0

One of the differences is that bootstrap does not preserve the order of the observations. It will sample observations (or blocks of observations, if using Block Bootstrap) in random order. Because you mentioned MLdP, I assume that you are working with financial time series. A possible disadvantage of bootstrapping is that it will not always preserve some ...


2

Bootstrap is not used to estimate mean. It can be used to estimate the uncertainty of the estimate for mean. You can use it in same way to estimate the uncertainty around any statistic, by following the same procedure: sampling with replacement from the data and evaluating the statistic on the sample, to eventually learn the distribution of the estimates by ...


Top 50 recent answers are included