|
Dear Group, I am a researcher of Linguistics, from India. I am trying to know Statistical Machine Learning so that it can be used appropriately in my extended field, Natural Language Processing. To learn this some time I get confused, how much I should know to handle a problem quite independently. Before starting this, I tried to read, (i)Mathematics esp. calculus; (ii)Statistics with an indepth knowledge of probability, distributions, etc. (iii) Neural Networks with a detail knowledge of Perceptron, MLP, etc. (iv) Algorithmics with details of data structure and all. (v) I am a fluent Python/C/C++ programmer. Is there any standard or level people in the community expects? What are the things I have to read/understand to gain a good confidence? If any one in the room can suggest me. Best Regards, Subhabrata Banerjee. |
|
The things you already read will be of great help. I am not very familiar with linguistics but a great book on statistical machine learning is bishop's "Machine Learning and Pattern Recognition". It is definitely worth buying. There is a book by Barber that is available online for free and seems to be quite good. Also, here is a blog post that gives many pointers to a wide variety of books in the field. |
|
Dear Sir, Thanks. I went through them. As your room suggested, I see I read quite good number of materials. All I need is slight confidence to handle the problems. Slightly new into the field so I feel once I would be able to handle few problems successfully I believe I would be in a confident position. So, I am trying to read some unique problems, see how they are doing and how I am feeling it, coding them seeing whether my thoughts are producing results. Wishing You A Happy Day Ahead, Best Regards, Subhabrata. |
|
Dear Sir, Thanks for your kind answer and kind words. Actually I am from a different background, a non Engineering. The way I interpret Machine Learning papers and devise algorithms, people around say me I am doing great. But, again sometimes I spend almost a day on some silly issues, like one day some of your colleague spent almost few hours teaching me posterior. So things get confusing sometimes. I have reviewed bishop's "Machine Learning and Pattern Recognition" really commendable one, but unfortunately not available in local book stores. I have just completed Barbers book and trying to do some exercises now. I am co-referring another book Data Mining Concepts and Techniques by Jiwaei Han and Micheline Kamber, the language of the book I found is very simple. I feel to get some more depth and confidence, I have to handle some more problems. Wishing You A Happy Day Ahead, Best Regards, Subhabrata. |
|
I first got started using the following: Assortment of tutorials: http://www.autonlab.org/tutorials/ Class from Berkeley on machine learning: http://www.cs.berkeley.edu/~jordan/courses/294-fall09/ |
|
Do check out the following: http://www.stanford.edu/class/cs229/materials.html and for the videos: http://www.youtube.com/playlist?p=A89DCFA6ADACE599 These are full lecture videos (with the accompanying set of notes) taught by Andrew. I took the class with him while an undergrad, he's a great teacher, very clear and really knows his stuff. |
|
Dear Sir,
Thanks for your kind update. I refer to Andrew Moore's Tutorials pretty often. The Berkeley is bit new to me. I will check the same.
Wishing You A Happy Day Ahead,
Best Regards,
Subhabrata. |
