ChanServ changed the topic of #mlpack to: "mlpack: a fast, flexible machine learning library :: We don't always respond instantly, but we will respond; please be patient :: Logs at
sourav has joined #mlpack
sourav has quit [Quit: Page closed]
Teshhfjjj has joined #mlpack
Teshhfjjj has quit [Client Quit]
Saurabh has joined #mlpack
Saurabh has quit [Quit: Page closed]
< davida> rcurtin: Nice pic - looks cold. I am suffering 27deg.C here. That's winter in Singapore ;)
Saurabh has joined #mlpack
< Saurabh> Hello Everybody !
< Saurabh> I am Saurabh Singh and I wish to contribute to MLPack. Is there any way out that I could contribute in Java? Please help me out
Saurabh has quit [Quit: Page closed]
Saurabh has joined #mlpack
Saurabh has quit [Client Quit]
< jenkins-mlpack2> Project docker mlpack nightly build build #144: STILL UNSTABLE in 7 hr 52 min:
davida has quit [Read error: Connection reset by peer]
davida has joined #mlpack
cult- has joined #mlpack
< rcurtin> davida: wow, 27C in December... too hot for me :)
< davida> ... and this is the cooler rainy season.
< rcurtin> whew
< rcurtin> well I guess you are right near the equator
< jenkins-mlpack2> Yippee, build fixed!
< jenkins-mlpack2> Project mlpack - git commit test build #60: FIXED in 37 min:
< jenkins-mlpack2> noreply: Merge pull request #1584 from rcurtin/instantiated-tree-metric
travis-ci has joined #mlpack
< travis-ci> mlpack/mlpack#5646 (master - 823c34d : Ryan Curtin): The build has errored.
travis-ci has left #mlpack []
< cult-> zoq: i am not sure if you recall tha arbitrary eigenvalue signs
< cult-> <zoq> cult-: The idea is relative simple:
< cult-> <zoq> arma::rowvec signs = arma::sign(arma::max(eigvec, 0));
< cult-> <zoq> eigvec.each_row() %= signs;
< cult-> <zoq> after the svd call in the exact svd method is basically all you need; just have to make sure I use the right operation each_row or each_col
< cult-> zoq: my question is 1) what if the eigenvector is a rolling eigenvector 2) did you figure out any solution since we last talked about this?
harkirat has joined #mlpack
harkirat has left #mlpack []
< zoq> cult-: I have opened a PR:, but couldn't backport the missing armadillo function yet.
ImQ009 has joined #mlpack
< cult-> zoq: incredible funny, I have just found the exact paper.
< cult-> =)
< cult-> zoq: awesome implementation!
< cult-> zoq: which function is missing?
< cult-> uhm, wait a minute
< cult-> "Common to the above examples is the lack of a fundamental technique for determining the sign.
< cult-> Therefore ad hoc approaches are used such as setting the maximum element in a singular vector to be
< cult-> positive. Unfortunately, such approaches do not solve the problem from a data analytical or
< cult-> interpretational point of view."
< cult-> indeed it doesn't solve the problem fundamentally. So how should one approach this?
davida has quit [Read error: Connection reset by peer]
davida has joined #mlpack
ImQ009 has quit [Quit: Leaving]
< rcurtin> cult-: based on the travis build log, it looks like index_max() is missing. The minimum version of armadillo that has them is 7.300.0, so if you have that Armadillo version or newer you can just patch the PR into your code and use it
< cult-> rcurtin: thanks