How to Study Machine Learning Practically | Coursera Community
Coursera Header
Solved

How to Study Machine Learning Practically

  • 9 April 2019
  • 7 replies
  • 285 views

Userlevel 1
Badge +1
I have taken a few courses I get all the points and at last I am confused where to use these all techniques. We have a lot of APIs to minimise our efforts sometimes i guess why I learned these . SO can you all suggest me me how to start working on real machine learning problems.

Aditya Raman

(Moved to Data Science forum by @Laura)
icon

Best answer by Elias_P 13 May 2019, 08:18

Good morning @adityaraman

You can start with Coursera's quiz: https://www.coursera.org/career/assessment/data-science

At the end of the 7 questions you'll get your answer for your career. Then, if you scroll down, you'll see the courses that are recommended for you. I assure you that I haven't find any course here at Coursera that sucks. At least not until today. So, go ahead and choose one!

At every course, you'll do, at least, one project. Don't stay to your instructor's solution. Try and make it better. 89% is good but 92% is better. And it will be YOUR achievement!

When you'll finish the Specialization try what you've learned on your project. Let's say that you want to do NLP. Find a book that you love, count the words, find the connection between the characters, visualize it...

The roadmap that works for me is: Study, play with what I've learned, study, experiment, study, explore and always but always have fun! If you don't have fun doing it then stop it. Change to something else. Only when you have fun, you'll love it, you'll spend enough time for it and you'll have time for yourself aso.

Please keep in mind that this is MY roadmap. Try and find yours. There isn't only one roadmap that will take you to your target.
View original

7 replies

Userlevel 7
Badge +13
@hamster @Kalyan @THANGA MANICKAM M @Elias_P – Do you have any suggestions for @adityaraman?
Userlevel 4
Badge +4
Hi @adityaraman ,

If you don't have something on your mind, like an everyday problem that you are facing and you want to solve it, then read the following article. You'll find great open source projects:

https://medium.mybridge.co/amazing-machine-learning-open-source-tools-projects-of-the-year-v-2019-95d772e4e985

Hope it'll help you!

Or you can play with Google Cloud Platform (GCP). You'll get a $300 on your account for a year (whatever finishes first). You can build the projects you've finished from your courses using the GCP. By doing this, you'll get hands on experience on both ML and GCP.

Isn't it great? 😉

Thanks for mentioning me @Laura 😎
Userlevel 1
Badge +1
Hi @adityaraman ,

If you don't have something on your mind, like an everyday problem that you are facing and you want to solve it, then read the following article. You'll find great open source projects:

https://medium.mybridge.co/amazing-machine-learning-open-source-tools-projects-of-the-year-v-2019-95d772e4e985
Hope it'll help you!

Or you can play with Google Cloud Platform (GCP). You'll get a $300 on your account for a year (whatever finishes first). You can build the projects you've finished from your courses using the GCP. By doing this, you'll get hands on experience on both ML and GCP.

Isn't it great? 😉


Thanks for mentioning me @Laura 😎

Thank you for this @Elias_P .
Can you suggest some roadmap to learn machine learning and data science. Like how to begin and in which order it should be completed in order to learn most of its parts to gain maximum knowledge.
Userlevel 4
Badge +4
Good morning @adityaraman

You can start with Coursera's quiz: https://www.coursera.org/career/assessment/data-science

At the end of the 7 questions you'll get your answer for your career. Then, if you scroll down, you'll see the courses that are recommended for you. I assure you that I haven't find any course here at Coursera that sucks. At least not until today. So, go ahead and choose one!

At every course, you'll do, at least, one project. Don't stay to your instructor's solution. Try and make it better. 89% is good but 92% is better. And it will be YOUR achievement!

When you'll finish the Specialization try what you've learned on your project. Let's say that you want to do NLP. Find a book that you love, count the words, find the connection between the characters, visualize it...

The roadmap that works for me is: Study, play with what I've learned, study, experiment, study, explore and always but always have fun! If you don't have fun doing it then stop it. Change to something else. Only when you have fun, you'll love it, you'll spend enough time for it and you'll have time for yourself aso.

Please keep in mind that this is MY roadmap. Try and find yours. There isn't only one roadmap that will take you to your target.
Userlevel 1
Badge +1
Thank You @Elias_P , your suggestions are really very nice and it helped me very much.
Thank you once again.
Userlevel 4
Badge +4
@adityaraman You are welcome! 😉
Badge
You can get one perspective on activity in the applied machine learning world by reviewing competitions at Kaggle. Some examples include earthquake prediction with Los Alamos National Laboratory (US), Customer transaction prediction (Santander), Histopathologic Cancer Detection, Malware Target prediction (Microsoft). One I looked into a while back was accurately classifying icebergs from X-band radar images for a company that operates oil rigs in the North Sea. These competitions typically provide datasets and quality metrics. Sometimes there are suggested approaches or starter kernels that can be reused from other competitions or competitors. Sometimes there are significant prizes for competition winners.

Reply