[sympy] Re: GSoC'19 Probability Idea proposal

2019-03-22 Thread Ritesh Kumar
*Implementation plans* STAGE 1 I plan to start the project by implementing the following distributions: - Univariate(Continuous): - Inverse Gaussian distribution: [https://en.wikipedia.org/wiki/Inverse_Gaussian_distribution] - Levy distribution: [https://en.wikipedi

[sympy] Re: GSoC'19 Probability Idea proposal

2019-03-22 Thread Ritesh Kumar
Here is my proposal. Please review and give your feedback. Regards *Implementation plans* STAGE 1 I plan to start the project by implementing the following distributions: - Univariate(Continuous): - Inverse Gaussian distribution: [https://en.wikipedia.org/wiki/Inverse_Gaussian_d

[sympy] Re: GSoC'19 Probability Idea proposal

2019-03-22 Thread Ritesh Kumar
STAGE 4 Export expressions of RV to external libraries(PyMC3, PyStan) I have used PyMC3 in my Machine Learning courses. Since both are python libraries, exporting to these libraries would basically deal with translating SymPy RVs so that they can be fed into them. Below is an example, to con

Re: [sympy] Re: GSoC'19 Probability Idea proposal

2019-03-23 Thread Aaron Meurer
I don't have much experience with the stats module, so others will need to comment on the details. It sounds like you have identified several things that it currently cannot do but could do. One suggestion would be to include the correct answer that SymPy should return for each of your examples whe

Re: [sympy] Re: GSoC'19 Probability Idea proposal

2019-03-23 Thread Ritesh Kumar
Thank you, Aaron, for your reply. I have made the corrections as you suggested. I was looking for the distributions in SymPy docs , and could not find Gumbel there. I will add Gumbel in the documentation and will also look for other distributions which

Re: [sympy] Re: GSoC'19 Probability Idea proposal

2019-03-25 Thread Aaron Meurer
Take a look at who the student and mentors were for the project in the stats module last year (or it may have been 2016, I can't remember). Aaron Meurer On Sat, Mar 23, 2019 at 1:27 PM Ritesh Kumar wrote: > Thank you, Aaron, for your reply. I have made the corrections as you > suggested. I was