[Scikit-learn-general] Decision Tree Classification of Raster Data using scikit-learn

Hi there, Can anyone help me out of this. I am using Scikit-learn module of Python for classification analysis of my data. I am using MODIS satellite sensor image (raster data in .hdf format having 1-7 spectral bands) and I actually need to perform decision tree classification of this dataset. I

Re: [Scikit-learn-general] Contributing with Kernel partial least squares

2013/12/5 abdalrahman eweiwi : > Hi, > > I have a problem of getting GridSearchCV working with kpls for fine tunning > its parameters. I have reviewed the documentation for how to get a new > estimator work with GridSearchCV but I did not find enough information > there. I pursued myself on this bu

Re: [Scikit-learn-general] Contributing with Kernel partial least squares

Hi, I have a problem of getting GridSearchCV working with kpls for fine tunning its parameters. I have reviewed the documentation for how to get a new estimator work with GridSearchCV but I did not find enough information there. I pursued myself on this but finally I got stuck with a problem relat

Re: [Scikit-learn-general] Find the pattern causing clustering in the data (if any)

2013/12/5 Luca Cerone : > I have a dataset made of N elements each which m features. Each item is > either > labelled as A or as B. Are there ways using scikit-learn to have some > clustering > algorithm that forces the clusters to be A and B, and returns the set of > rules > that determine such cl

[Scikit-learn-general] Find the pattern causing clustering in the data (if any)

Dear all, I would like to know how to solve this problem I have using scikit-learn. I have a dataset made of N elements each which m features. Each item is either labelled as A or as B. Are there ways using scikit-learn to have some clustering algorithm that forces the clusters to be A and B, and

Re: [Scikit-learn-general] Grid Search on Multiple scoring functions

2013/12/5 Joel Nothman : > This is a frequent request, and hopefully we will soon find consensus on an > API for it. Please see > https://github.com/scikit-learn/scikit-learn/issues/1850 and offer any > suggestions! The trick for now is to define a custom scorer that returns the score you're optim

Re: [Scikit-learn-general] Grid Search on Multiple scoring functions

Hi Karthick, This is a frequent request, and hopefully we will soon find consensus on an API for it. Please see https://github.com/scikit-learn/scikit-learn/issues/1850 and offer any suggestions! Cheers, Joel On Thu, Dec 5, 2013 at 1:17 AM, Karthick M wrote: > Hi, > Is there any way to do Gr

[Scikit-learn-general] efficient way to classifier predict

Hi there , I am doing svm classification for remote sensing images usisg scikit learn. My problem is that for image size of (2000,2000,4), the input to the clf.predict is of size 2000x2000,4 i,e four millions rows and clf.predict is taking forever (26 min) to give me the output. Is there an effiec

[Scikit-learn-general] Grid Search on Multiple scoring functions

Hi, Is there any way to do GridSearch on multiple scoring functions without having to repeat the fit/predict process for each function? (e.g.: grid_search_digits.py - In this example, GridSearch is run twice, once each for precision and recall. ) Possible Solution: Manual approach. Use IterGrid a