Re: [BangPypers] Resource for ML
I can vouch for Coursera's ML courses by University of Washington. It gives you a brief overview of the possibilities ML presents in the foundations course - predictive models using regression, document classification, recommender systems in the very first course - good for whetting your appetite with a black box approach. The next courses delve deep into each example, be it Regression, or Classification - with generous doses of math(some of it optional). They use jupyter notebooks and have adapted to present solutions in scikit-learn apart from the proprietary stuff they initially started off with because one of the instructors was a founder of an AI and ML start-up that they tried promoting ( Maybe you've heard of Dato, now goes by the name of Turi - acquired by Apple). It should give you a good, firm grasp on the basics and enough to keep you busy for a while. On Wed, Jun 7, 2017 at 7:41 PM, Anand Chitipothuwrote: > On Tue, Jun 6, 2017 at 8:59 PM, Ramkrishna P > wrote: > > > Hello Team, > > I have started out to work on pandas and numpy libraries to pick some > > machine learning concepts. > > I feel apart from working on datasets and getting some results, the > > core concepts of machine learning are still missing. > > > > If you guys could suggest some resources, it will be of great help. > > > > I find Learn Data Science very good place to start, esp. for the beginners. > > http://learnds.com/ > > Anand > ___ > BangPypers mailing list > BangPypers@python.org > https://mail.python.org/mailman/listinfo/bangpypers > -- Arjunil Pathak ___ BangPypers mailing list BangPypers@python.org https://mail.python.org/mailman/listinfo/bangpypers
Re: [BangPypers] Resource for ML
On Tue, Jun 6, 2017 at 8:59 PM, Ramkrishna Pwrote: > Hello Team, > I have started out to work on pandas and numpy libraries to pick some > machine learning concepts. > I feel apart from working on datasets and getting some results, the > core concepts of machine learning are still missing. > > If you guys could suggest some resources, it will be of great help. > I find Learn Data Science very good place to start, esp. for the beginners. http://learnds.com/ Anand ___ BangPypers mailing list BangPypers@python.org https://mail.python.org/mailman/listinfo/bangpypers
Re: [BangPypers] Resource for ML
Hello, while not having finished Andrew Ng's coursera course (yet), I started it and like it, too. I don't think it's an disadvantage that it's Matlab (or it's open source counterpart, Octave) - based (and I'm much more proficient in Python than in Matlab). Thanks to Abhinav and Harsh for the other recommendations. Cheers, Nenad 2017-06-06 17:44 GMT+02:00 Abhinav Upadhyay: > On Tue, Jun 6, 2017 at 8:59 PM, Ramkrishna P > wrote: > > Hello Team, > > I have started out to work on pandas and numpy libraries to pick some > > machine learning concepts. > > I feel apart from working on datasets and getting some results, the > > core concepts of machine learning are still missing. > > > > If you guys could suggest some resources, it will be of great help. > > Andrew Ng's coursera course is probably the best place to start, he > covers a broad range of models which are commonly used and builds > mathematical intuitions for each of them (without bogging you down > with proofs, which have their place but not at this stage). Although, > all the programming exercises in the course use GNU Octave or Matlab. > > For a slightly more in depth coverage, you may consider the University > of Washington's specialization on ML (available on Coursera). It is a > set of 4 courses. The first course is just dedicated to regression, > while the second one just covers classification models. So every > course is able to go into more details than Ng's course. As a bonus, > all the exercises in the courses use Python. > > For a more statistics oriented introduction there is a course on > Stanford Online from Trevor Hastie and Rob Tibshirani based on their > book Introduction to Statistical Learning. All the exercises use R. > > PS: All the courses can be easily found with the help of Google, I > didn't have the links handy. > > - > Abhinav > ___ > BangPypers mailing list > BangPypers@python.org > https://mail.python.org/mailman/listinfo/bangpypers > ___ BangPypers mailing list BangPypers@python.org https://mail.python.org/mailman/listinfo/bangpypers
Re: [BangPypers] Resource for ML
I collected some ML resources for inter hostel data analytic competition here https://github.com/Azad-Hall/data-analytics Other the Andrew ng's course, Caltech's "Learning from Data" ( http://work.caltech.edu/telecourse.html) course is really good for the theoretical foundations of ML> On 6 June 2017 at 21:14, Abhinav Upadhyaywrote: > On Tue, Jun 6, 2017 at 8:59 PM, Ramkrishna P > wrote: > > Hello Team, > > I have started out to work on pandas and numpy libraries to pick some > > machine learning concepts. > > I feel apart from working on datasets and getting some results, the > > core concepts of machine learning are still missing. > > > > If you guys could suggest some resources, it will be of great help. > > Andrew Ng's coursera course is probably the best place to start, he > covers a broad range of models which are commonly used and builds > mathematical intuitions for each of them (without bogging you down > with proofs, which have their place but not at this stage). Although, > all the programming exercises in the course use GNU Octave or Matlab. > > For a slightly more in depth coverage, you may consider the University > of Washington's specialization on ML (available on Coursera). It is a > set of 4 courses. The first course is just dedicated to regression, > while the second one just covers classification models. So every > course is able to go into more details than Ng's course. As a bonus, > all the exercises in the courses use Python. > > For a more statistics oriented introduction there is a course on > Stanford Online from Trevor Hastie and Rob Tibshirani based on their > book Introduction to Statistical Learning. All the exercises use R. > > PS: All the courses can be easily found with the help of Google, I > didn't have the links handy. > > - > Abhinav > ___ > BangPypers mailing list > BangPypers@python.org > https://mail.python.org/mailman/listinfo/bangpypers > -- Harsh Sent from a GNU/Linux ___ BangPypers mailing list BangPypers@python.org https://mail.python.org/mailman/listinfo/bangpypers
Re: [BangPypers] Resource for ML
Hey Ramkrishna, I have found the following book very useful. - https://github.com/jakevdp/PythonDataScienceHandbook Thank you, Bhargav On Tue, Jun 6, 2017 at 8:59 PM, Ramkrishna Pwrote: > Hello Team, > I have started out to work on pandas and numpy libraries to pick some > machine learning concepts. > I feel apart from working on datasets and getting some results, the > core concepts of machine learning are still missing. > > If you guys could suggest some resources, it will be of great help. > > > Regards, > Ramkrishna.P > ___ > BangPypers mailing list > BangPypers@python.org > https://mail.python.org/mailman/listinfo/bangpypers > ___ BangPypers mailing list BangPypers@python.org https://mail.python.org/mailman/listinfo/bangpypers