Mailman 3 March 2017 - scikit-learn

Using Scikit-Learn to predict magnetism in chemical systems
by Henrique C. S. Junior March 28, 2017

March 28, 2017

I'm a chemist with some rudimentary programming skills (getting started with python) and in the middle of the year I'll be starting a Ph.D. project that uses computers to describe magnetism in molecular systems. Most of the time I get my results after several simulations and experiments, so, I know that one of the hardest tasks in molecular magnetism is to predict the nature of magnetic interactions. That's why I'll try to tackle this problem with Machine Learning (because such interactions are dependent, basically, of distances, angles and number of unpaired electrons). The idea is to feed the computer with a large training set (with number of unpaired electrons, XYZ coordinates of each molecule and experimental magnetic couplings) and see if it can predict the magnetic couplings (J(AB)) of new systems: (see example in the attached image) Can Scikit-Learn handle the task, knowing that the matrix used to represent atomic coordinates will probably have a different number of atoms (because some molecules have more atoms than others)? Or is this a job better suited for another software/approach? -- *Henrique C. S. Junior* Industrial Chemist - UFRRJ M. Sc. Inorganic Chemistry - UFRRJ Data Processing Center - PMP Visite o Mundo Químico <http://mundoquimico.com.br>

4 8

Create sample dataset with specified range and regression coefficients
by Onkar Mahajan March 27, 2017

March 27, 2017

I would like to create a random dataset for Linear regression with specific regression coefficients (reg.coef_, reg.intercept_ ) and with data in specific range of values (person age - 0 to 100 as x - axis and y-axis Net worth 0$ to 5,00,000$). I used sklearn.datasets.make_regression() but I did not find anything in here that gives me control over range of samples and regression coefficients (I might be missing something, please correct me if mistaken) Thanks, Onkar

2 1

GSoC 2017 : "Parallel Decision Tree Building"
by Aman Pratik March 27, 2017

March 27, 2017

Hello Developers, This is Aman Pratik. I am currently pursuing my B.Tech from Indian Institute of Technology, Varanasi. After doing some research I have found some material on Decision Trees and Parallelization. Hence, I propose my first draft for the project "Parallel Decision Tree Building" for GSoC 2017. Proposal : First Draft <https://github.com/amanp10/scikit-learn/wiki/GSoC-2017-:-Parallel-Decision-…> Why me? I have been working in Python for the past 2 years and have good idea about Machine Learning algorithms. I am quite familiar with scikit-learn both as a user and a developer. These are the issues/PRs I have worked/working on for the past few months. [MRG+1] Issue#5803 : Regression Test added #8112 <https://github.com/scikit-learn/scikit-learn/pull/8112> [MRG] Issue#6673:Make a wrapper around functions that score an individual feature #8038 <https://github.com/scikit-learn/scikit-learn/pull/8038> [MRG] Issue #7987: Embarrassingly parallel "n_restarts_optimizer" in GaussianProcessRegressor #7997 <https://github.com/scikit-learn/scikit-learn/pull/7997> My GitHub Profile: amanp10 <https://www.github.com/amanp10> I have worked with parallelization in one of my PR, so I am not new to it. I have used cython a couple of times, though as a beginner. I have not used Decision Tree much but I am familiar with the theory and algorithm. Also, I am familiar with Benchmark tests, Unit tests and other technical knowledge I would require for this project. Meanwhile, I have started my study for the subject and gaining experience with Cython. I am looking forward to guidance from the potential mentors or anyone willing to help. Thank You

2 4

Preparing a scikit-learn 0.18.2 bugfix release
by Olivier Grisel March 26, 2017

March 26, 2017

Hi all, I think we should release 0.18.2 to get some important fixes and make it easy to release Python 3.6 wheel package for all the operating systems using the automated procedure. I identified a couple of PR to backport to 0.18.X to prepare the 0.18.2 release. Are there any other important recently fixed bugfs people would like to see backported in this release? https://github.com/scikit-learn/scikit-learn/milestone/23?closed=1 Best, -- Olivier http://twitter.com/ogrisel - http://github.com/ogrisel

5 16

Note of appreciation to Scikit-learn team
by Brown J.B. March 23, 2017

March 23, 2017

To all organizers, developers, and maintainers involved in the Scikit-learn project, I would like to share a recent article that researchers from MIT, ETH, and Kyoto University (myself) have published about building efficient models for drug discovery and pharmaceutical data mining. In short, it demonstrates through replicate experiment that neither big data nor complex AI such as deep learning are necessary for efficient drug discovery, and that active learning can guide/assist decision making processes in the real world. The paper's success is underpinned by the use of Scikit-learn's RandomForestClassifier implementation combined with other techniques developed in the work. Therefore, it is a by-product of the volunteerism, hard work, and dedication by those involved in scikit-learn. As the senior author of this study, I wish to share my great appreciation for your efforts. While I am strongly limited in time and can barely contribute to this community, I cannot thank all of you enough for your work - it has made an impact. We are working on theoretical extensions of the work now, as well as pushing the technology forward in applied discovery sciences (in agricultural, pharmaceutical, and medical areas). In the theory and real-world applications, scikit-learn is indispensible. We have made the paper open access, and hope that such will inspire this community as well as those in applied sciences. You will see that the open source software community has been listed in the Acknowledgments. Certainly, we would welcome even the most casual of comments about the paper. The paper can be retrieved from here: http://www.future-science.com/doi/abs/10.4155/fmc-2016-0197 With kindest regards and sincere appreciation, J.B. Brown Kyoto University Graduate School of Medicine Junior Associate Professor and Principal Investigator http://statlsi.med.kyoto-u.ac.jp/~jbbrown PS - To those of you involved in the matplotlib, scipy, and numpy projects, your forwarding of this to those projects would be appreciated. They were also critical.

4 3

GSoC proposal - Improve online learning for linear models
by Konstantinos Katrioplas March 23, 2017

March 23, 2017

Hello all, Please review my proposal on improving the online learning for linear models: first draft - linear model proposal <https://docs.google.com/document/d/1mRj-yxxhtapKGqFdA6M6e7PG6-klm0-1VJGBnFM…> Please bear in mind that this is a first approach only. I would like your opinion on if this goes into the right direction and how it can be improved. On the tool to set the learning rate in particular, I need your ideas on how it could be implemented. Previously mentioned ideas on a callback function are interesting, but I would need some guidance on implementing that. Although I am interested in the decision trees as well, I feel the linear model is a better start for me as my intention is to keep contributing to scikit-learn after the summer. I have a background in computational physics, however I am much more focused on the computational side than the physics side. Here is my resume <https://drive.google.com/file/d/0BwZy58HBIWp7U0I1RDNyRGhsNUU/view?usp=shari…> . PRs and Issues I have been involved in so far: [MRG] enhance make_blobs to accept lists for samples per cluster <https://github.com/scikit-learn/scikit-learn/pull/8563> [MRG] add random_state in tests estimators <https://github.com/scikit-learn/scikit-learn/pull/8563> Bug in bfgs gradient computation of MLPRegressor with multiple output neurons <https://github.com/scikit-learn/scikit-learn/issues/8349> (I am very curious about this one) github profile: kkatrio <https://github.com/kkatrio> I am looking forward to your opinion. Kind regards, Konstantinos Katrioplas

1 0

Regarding GSoC projects and mentors
by Jeff Lee March 22, 2017

March 22, 2017

Hi, My name is Jefferson Lee and I am a computer science student at NYU passionate about machine learning and AI. I was hoping to speak with potential mentors regarding the two suggested projects <https://github.com/scikit-learn/scikit-learn/wiki/Google-summer-of-code-(GS…> and how I might research more about the topics to write a very strong proposal. I also wanted to gain learn about some of the stalled projects that might be viable as a GSoC project. Let me know if anyone is available to chat about these topics! Best, Jeff

3 2

GSoc, 2017 (proposal idea and intro) .reg
by SHUBHAM BHARDWAJ 15BCE0704 March 22, 2017

March 22, 2017

Hello Sir, My introduction : I am a 2nd year student studying Computer Science and engineering from VIT, Vellore. I work in Google Developers Group VIT. All my experience has been about collaborating with a lot of people ,working as a team, building products and learning along the way. Since scikit-learn is participating this time I am too planning to submit a proposal. Proposal idea: I am really interested in implementing kmeans++ algorithm.I was doing some work on DT but I found this very appealing. Just wanted to know if it can be a good project idea. Regards Shubham Bhardwaj

5 9

A custom loss function for GradientBoostingRegressor
by Zygmunt Zając March 20, 2017

March 20, 2017

Hello, I would like to add a custom loss function for gradient boosting regression. The function is similar to least squares, except that for each example it is OK to either undershoot or overshoot the target - loss is zero then. There is an additional binary indicator called "under" telling us whether it is OK to undershoot or overshoot. For example: y under p loss 5 1 4 0 5 0 4 1 5 1 6 1 5 0 6 0 Below is my attempt at implementation. I have three questions: 1. Is it correct? 2. How would you pass "under" to the loss function? 3. Functions other than LeastSquaresError() seem to _update_terminal_regions_. Is this necessary in this case, and if so, how to do it? def __call__(self, y, pred, sample_weight=None): if sample_weight is None: squares = (y - pred.ravel()) ** 2.0 # the custom part overshoot_ok = (pred > y) & (under == 0) undershoot_ok = (pred < y) & (under == 1) squares[overshoot_ok] = 0 squares[undershoot_ok] = 0 return np.mean(squares) else: (...) def negative_gradient(self, y, pred, **kargs): diffs = y - pred.ravel() overshoot_ok = (pred > y) & (under == 0) undershoot_ok = (pred < y) & (under == 1) diffs[overshoot_ok] = 0 diffs[undershoot_ok] = 0 return diffs

1 0

Please unsubscribe
by Stratford, Mark A March 20, 2017

March 20, 2017

-----Original Message----- From: scikit-learn [mailto:scikit-learn-bounces+mark_stratford=optum.com@python.org] On Behalf Of scikit-learn-request(a)python.org Sent: Monday, March 20, 2017 6:06 AM To: scikit-learn(a)python.org Subject: scikit-learn Digest, Vol 12, Issue 42 Send scikit-learn mailing list submissions to scikit-learn(a)python.org To subscribe or unsubscribe via the World Wide Web, visit https://mail.python.org/mailman/listinfo/scikit-learn or, via email, send a message with subject or body 'help' to scikit-learn-request(a)python.org You can reach the person managing the list at scikit-learn-owner(a)python.org When replying, please edit your Subject line so it is more specific than "Re: Contents of scikit-learn digest..." Today's Topics: 1. recommended feature selection method to train an MLPRegressor (Thomas Evangelidis) 2. Re: recommended feature selection method to train an MLPRegressor (Andreas Mueller) 3. Re: recommended feature selection method to train an MLPRegressor (Sebastian Raschka) 4. Anomaly/Outlier detection based on user access for a large application (John Doe) ---------------------------------------------------------------------- Message: 1 Date: Sun, 19 Mar 2017 20:47:36 +0100 From: Thomas Evangelidis <tevang3(a)gmail.com> To: Scikit-learn user and developer mailing list <scikit-learn(a)python.org> Subject: [scikit-learn] recommended feature selection method to train an MLPRegressor Message-ID: <CAACvdx17Ev3jr0ds2bLyJc0RqZkqJH7Rtx=s1ZaodmUvCkcB8Q(a)mail.gmail.com> Content-Type: text/plain; charset="utf-8" Which of the following methods would you recommend to select good features (<=50) from a set of 534 features in order to train a MLPregressor? Please take into account that the datasets I use for training are small. http://scikit-learn.org/stable/modules/feature_selection.html And please don't tell me to use a neural network that supports the dropout or any other algorithm for feature elimination. This is not applicable in my case because I want to know the best 50 features in order to append them to other types of feature that I am confident that are important. ?cheers Thomas? -- ====================================================================== Thomas Evangelidis Research Specialist CEITEC - Central European Institute of Technology Masaryk University Kamenice 5/A35/1S081, 62500 Brno, Czech Republic email: tevang(a)pharm.uoa.gr tevang3(a)gmail.com website: https://sites.google.com/site/thomasevangelidishomepage/

1 0