verne.freenode.net changed the topic of #mlpack to: http://www.mlpack.org/ -- We don't respond instantly... but we will respond. Give it a few minutes. Or hours. -- Channel logs: http://www.mlpack.org/irc/
stephent1 has quit [Ping timeout: 246 seconds]
awhitesong has joined #mlpack
awhitesong has left #mlpack []
curiousguy13 has quit [Ping timeout: 264 seconds]
curiousguy13 has joined #mlpack
curiousguy13 has quit [Ping timeout: 250 seconds]
JaraxussTong has quit [Quit: No Ping reply in 180 seconds.]
JaraxussTong has joined #mlpack
stephentu has quit [Quit: Lost terminal]
curiousguy13 has joined #mlpack
curiousguy13 has quit [Ping timeout: 255 seconds]
curiousguy13 has joined #mlpack
curiousguy13 has quit [Ping timeout: 264 seconds]
kshitijk has joined #mlpack
vlad_gl has joined #mlpack
vlad_gl has quit [Ping timeout: 246 seconds]
stephentu has joined #mlpack
vlad_gl has joined #mlpack
kshitijk has quit [Ping timeout: 264 seconds]
< vlad_gl> hi! i start digging gmm method, and look at #215 task now. i want try to build binaryspace tree on gaussians means using mahalanobius distance as metric.
< vlad_gl> after i do this, what is the correct way to find nearest neighbour for any observation?
< vlad_gl> actually, i not fully looked trees interface now..
< vlad_gl> anyway, is it right solution?
< naywhayare> hi vlad, I think one problem is that the bound type for BinarySpaceTree (HRectBound<>) only works with LMetric, so I am not sure that you can use the Mahalanobis distance
< naywhayare> but you could use the cover tree:
< naywhayare> CoverTree<MahalanobisDistance> tree(means); // means is the matrix where each column is an observation
< naywhayare> or just use the NeighborSearch object:
< naywhayare> NeighborSearch<NearestNeighborSort, MahalanobisDistance, CoverTree<MahalanobisDistance> > knn(queryPoints, means); // query points are the points for which you want the nearest mean
< naywhayare> then knn.Search(...)
< vlad_gl> ok, i
< vlad_gl> i'll try it. thanks :)
stephentu has quit [Ping timeout: 252 seconds]
< naywhayare> vlad_gl: here is an interesting paper that may be relevant: http://www.ri.cmu.edu/pub_files/pub1/moore_andrew_1998_2/moore_andrew_1998_2.pdf
< naywhayare> I don't know how well it's written since I haven't taken the time to read it, but the basic idea is that you can use trees during EM
< naywhayare> I think this is pretty similar to what you're doing
< vlad_gl> yeah, looks like what I was looking for. thanks.
vlad_gl has quit [Ping timeout: 246 seconds]
kshitijk has joined #mlpack
stephentu has joined #mlpack
kshitijk has quit [Ping timeout: 246 seconds]
stephentu has quit [Ping timeout: 246 seconds]
stephentu has joined #mlpack
< naywhayare> stephentu: do you want to make an account on the Melange site for GSoC so I can add you as a potential mentor?
< stephentu> naywhayare: sure
< stephentu> naywhayare: ok I am registered as stephentu
< naywhayare> okay, I sent the invitation... I think it sends you an email
< stephentu> ok i think i accepted, did it work?
< naywhayare> yeah, it lists you as a mentor now
< stephentu> great
< stephentu> i'll write upa few more proposals
stephentu has quit [Ping timeout: 246 seconds]
JaraxussTong has quit [Quit: No Ping reply in 180 seconds.]
JaraxussTong has joined #mlpack
stephentu has joined #mlpack