There are two best times to plant a tree: ten years ago and now. This series is a summary of me, as a layman of machine learning, when I took the machine learning course of Professor Wu Enda of Stanford University. I have learned the fifth week. I will share some things in the learning process, hoping to help machine learning. Those who are interested and want to learn by themselves.
If you are interested in machine learning, but that's it, you haven't gotten started yet, and you haven't learned it systematically. Then this course is suitable for you. It is an introductory course. All the things you need, such as a little knowledge of linear algebra and the use of programming tools Octave/Matlab, have corresponding courses and exercises in the video to help you master. This is widely recognized as the best course to enter the field of machine learning, 1.8 million people have taken this course , and this course is also the most popular online course for 17 years .
Many people may have questions when learning machine learning by themselves, is this course suitable for them? What can you get? What conditions are required? Like Python or Linear Algebra? Abandon these doubts, if you want to learn, learn it, learn first, you will have a feeling, and your intuition will guide you.
To quote a passage from the great god Carmack in <<Doom Apocalypse>>:
In the information age, objective obstacles no longer exist, so-called obstacles are subjective. If you want to start developing something new, you don't need millions of dollars, you just need a refrigerator full of pizza and Coke, a cheap computer, and the determination to devote yourself to it. We slept on the floor, we waded through the river.
When you are studying this course, in addition to the video course, there are the following supporting resources that can be used to improve the efficiency of learning
The video was taken a few years ago (11 years? It is said to be the first course when coursera was launched), there may be some mistakes in it, pay attention to the errata:
such as Errata Week3Additional test cases for programming assignments Since there are only one or two test cases in the programming practice questions, there may be cases where you find that the problem is not covered after you submit it. In this case, you can use these additional test cases to help you find the problem.
FAQ for weekly courses and programming exercises: For
example , in the third week,
you can browse the content here before watching the video course, and then browse the content here after you finish the video course.Weekly focus summary
This is a discussion initiated by Professor Wu Enda. The main purpose is to let everyone retell the 5 key points of this week in their own words. Only what you can retell is what you really learn. A highly condensed summary. The problem with video courses is that the content they carry is limited, and it is easy to follow the video and lose the time to think.Study group , there are slack, what's app and WeChat group
I also set up a WeChat group , suitable for students who started studying recently (April 2018)Weekly related discussion groups, such as week three .
You can see some interesting ideas from it, such as drawing the cost function of linear regression and logistic regression, a picture is worth a thousand words; or some people will show the eigenvectors given in the topic of logistic regression in the fourth week, which is also very interesting.
References:
Recommended Courses and Study Groups for Data Scientist or Machine Learning Engineer
Kaggle: competitions, datasets, kernels
Youtube video of Yann Lecun, the father of Convolutional Neural Networks (CNN)
Youtube video of Jeffrey Hinton, godfather of artificial neural network (ANN)
10 Most Popular Courses in 2017
If you found this article helpful, please click the "Recommend" button to let more people see it!
