🌐 Mailman 3 September 2016 - scikit-learn - python.org

Fwd: Loading file in libsvm format
by klo uo Sept. 8, 2016

Sept. 8, 2016

---------- Forwarded message ---------- From: klo uo <klonuo(a)gmail.com> Date: Thu, Sep 8, 2016 at 8:25 PM Subject: Loading file in libsvm format To: scikit-learn-general(a)lists.sourceforge.net Hi, I produced a file in libsvm format: <label> <index1>:<value1> <index2>:<value2> ... with this content: 6284 576:1 884:1 2482:1 4279:1 5765:1 184552:1 661512:1 699842:1 2259 1669:1 5711528:6 2822 5765159:1 ... The label is document_id, and index:value are term_id and term count. This file has 83K labels with 40K unique terms (and overall 1.2M index:value pairs). When I load this file in sklearn: from sklearn.datasets import load_svmlight_file X, y = load_svmlight_file('libsim.txt') I get X with shape (82448, 6092168). I don't know of any reason why am I getting 6M features? Can someone explain? Thanks

2 2

Mailing list "slow"?
by Sebastian Raschka Sept. 8, 2016

Sept. 8, 2016

Hi, all, I noticed that it takes forever now until something is posted on the mailing list after I sent it out. Since the switch to Python.org, it takes about ~15 - 45min after hitting “sent”. I’ve noticed this for months now and was wondering if this is normal or of there’s something going on with my particular mailing list account? (besides the mailing list, my usually arrives within 1-2 seconds, so it’s not a problem with my email client or server in general). Best, Sebastian

5 7

Tuning custom parameters using grid_search
by Piotr Bialecki Sept. 7, 2016

Sept. 7, 2016

Hi all, I am currently tuning some parameters of my xgboost model using scikit's grid_search, e.g.: param_test1 = {'max_depth':range(3,10,2), 'min_child_weight':range(1,6,2) } gsearch1 = GridSearchCV(estimator = XGBClassifier(learning_rate =0.1, n_estimators=762, max_depth=5, min_child_weight=1, gamma=0, subsample=0.8, colsample_bytree=0.8, objective= 'binary:logistic', nthread=4, scale_pos_weight=1, seed=2809), param_grid = param_test1, scoring='roc_auc', n_jobs=6, iid=False, cv=5) Before that I preprocessed my dataset X with some different methods. These preprocessing steps have some parameters too, which I would like to tune. I know that it is possible to tune the parameters of the preprocessing steps, if they are part pf my pipeline. E.g. if I am using PCA, I could tune the parameter n_components, right? But what if I have some "custom" preprocessing code with some parameters? Is it possible to create a scikit-compatible "object" of my custom code in order to tune the parameters in the pipeline with grid search? Imagine I would like to write a custom method FeatureMultiplier() with a parameter multiplier_value. Is it possible to create a scikit-compatible class out of this method and tune it with grid search? I thought I saw a talk about exactly this topic at some PyData in 2016 or 2015, but unfortunately I cannot find the video of it. Maybe I misunderstood the presentation at that time. Best regards, Piotr

3 3

How to get the factorization from NMF in scikit learn
by Raphael C Sept. 7, 2016

Sept. 7, 2016

I am trying to use NMF from scikit learn. Given a matrix A this should give me a factorization into matrices W and H so that WH is approximately equal to A. As a sanity check I tried the following: from sklearn.decomposition import NMF import numpy as np A = np.array([[0,1,0],[1,0,1],[1,1,0]]) nmf = NMF(n_components=3, init='random', random_state=0) print nmf.components_ This gives me a single 3 by 3 matrix as output. What is this representing? I want the two matrices W and H from the factorization. How can I get these two matrices? I am sure I am just missing something simple. Raphael

2 1

Confidence Estimation for Regressor Predictions
by Daniel Seeliger Sept. 2, 2016

Sept. 2, 2016

Dear all, For classifiers I make use of the predict_proba method to compute a Gini coefficient or entropy to get an estimate of how "sure" the model is about an individual prediction. Is there anything similar I could use for regression models? I guess for a RandomForest I could simply use the indiviual predictions of each tree in clf.estimators_ and compute a standard deviation but I guess this is not a generic approach I can use for other regressors like the GradientBoostingRegressor or a SVR. Thanks a lot for your help, Daniel

5 8

Declaring numpy and scipy dependencies?
by Andreas Mueller Sept. 1, 2016

Sept. 1, 2016

Hi all. I think since the pipy ecosystem improved a lot, we should properly declare the scipy and numpy dependencies, so that ``pip install scikit-learn`` works properly. The argument why we did not do this previously was that this would try to download and build numpy and scipy, which are pretty much guaranteed to fail / result in a very slow installation. This is no longer true. Any opinions? Andy

5 10