Видео

Great video! Thanks Monica

No you wasted your money. LOl. His course is a scam.

Good explanation Monica 👍

Congratulations.

How about engineer?🤔

the course that is on youtube is it the same one that is on coursera ?

first comment!! great video!!

This is Litterally the best advice I’ve received

I'm not going to lie it's pretty f****** easy for me to get a job. I go for the challenging stuff like plumbing and repair type work.

If dude applied to 735 jobs, then he is the problem.

Finding coding boring and having a degree in this field is like being a doctor when you are a misanthrope. Just shows how many got into this field because of money and the status.

it's easier to get a 🔫

I was doing IBM data science machine learning which was straight, easy and understandable structure but when I tried to go through stanford it was a bit difficult, although I am doing Master's Studies and have good continuous academic touch but stanford looks me complicated I don't know why. Instead of making difficult methodology there should be given priority to make people understood by making easy stuff.

Sounds like a project I'd get an F- 😂 Good explanation though!

Andrew Ng's course is one I recommend to students who struggle in some data science classes. As to the comments saying something like "it's too basic, you cannot land a job with this class": The point of Andrew Ng's class is to give a decent foundation and it does that really quite well. If you have a good foundation, you can build upon it with relative ease if and when you have to. If you don't, you'll struggle way more to acquire the skills later. And on top of this, you guys would be surprised to see how many "data science" jobs out there are relatively low-level in terms of required skills. Not every industry job will require you to know and use the latest and greatest modeling approaches. So it really depends on the sector and your role.

Got curious on this, does it teach you on how to build recommendation systems for example ?

No way it can land some one a job in 2024, this course is too basic

Thanks I am also taking this course and I completely agree with you - from on India

amazing ❤

Nice work Monica, here are some of my thoughts on your work: - The implementation of correlation in numpy is Pearson correlation and as you may know it measures the "linear association" of a pair of "continuous variables" which is not a correct measure of correlation between your target variable (continuous) and gender (binary). Instead you should use Point-Biserial Correlation. - Pearson correlation assumes normally distributed continuous variables which is not the case for the majority of independent variables. Additionally, Pearson correlation is affected by outliers if you take a look at formula so ideally you need to perform additional analysis to detect outliers in your data and decide what to do with it (remove/winsorize/scale). - All the algorithm you used in your project are intended to capture non-linear relationship between target and explanatory variables, not sure of the choice of Pearson correlation since it is a measure of linear association. Spearman rank correlation is a better candidate for non linear association measure since it captures monotonic relationships. - You can still keep the gender by encoding (one hot encoding) the variable (for random forest and AdaBoost) or converting the values of it into 2 dummy variables as for the case of polynomial regression, provided that there is a strong association between target variable and gender from point biserial test. - What is the point of checking for the normality of distribution of explanatory variables when you used models which do not strictly require normally distributed variables? - You can have correlated variables (not to the extreme of 0.9+) as final selected features since all the algorithms that you use have regularization to control for multicollinearity. - Polynomial regression model is sensitive to outliers, you probably need to have some sorts of treatment for outliers. - Since polynomial regression is a parametric model, it relies on several assumptions such as Independence of Errors, Homoscedasticity of the residual variance. Normality of Residuals, etc. I think this is something you can check cus if these assumptions do not hold, the results aren't statiscally reliable.

I was trying to find someone going through a deep review like this! I’m a current CS student and wanted to reinforce some of the topics I’ve learned already. Thank you so much! Can’t wait to start!

Quite interesting!! I'll be very grateful if you can provide me some guidance as I'm ultra pro Max beginner. 😭

Thank you for making this video! I just started this course. Hoping it will help me for getting ml role 😅

What kind of models are you working with at work?

How you landed the job . The course only covers basics of ML

your youtube channel is pretty good, but i think the SEO is not set up perfectly in your youtube videos, thats why the views is low, i think i can help you, let me know

programming careers are the best

I have almost completed the first course, What other resources should I follow along with these courses?

amazing 🎉

🤔 your video made me consider pursuing more knowledge in the machine learning area, seems to be quite useful in the modern era

amazing！

coding itself can indeed get boring and repetitive, but the result makes it worth it code what you like, and you end up with positive experience, code what you need, and you end up with bland experience, code anything else, and you end up with miserable experience,

If you think programmers will be replaced by an advanced Mr clippy you've been listening to the same doom salesman as those that peddle nuclear fusion, quantum computing and a hundred other scams

Great video!! I really enjoy exploring different viewpoints

Actually writing basic software functions is and always has been trivial. Software engineers are hired to design and maintain complex systems using software to meet customer needs. There is a large gap between that and generative AI for now and likely far into the future. If AI gets smart enough to replace all aspects of software design and maintenance, it will replace every single thought based job. If it gets so good that it can eventually just make an engineer produce products 10x faster, then we will see an explosion of productivity and products.

😿

My take is that the field is just getting even more competitive. I don't really think AI will replace programmers, but will surely translate the work that juniors before did, to the semi seniors o seniors in the industry. Heck, and not even yet, this friday I was helping a friend at work to use pagination in a website, he was using the paid version of chat gpt4, not only the strategies provided by the AI were wrong and not being used in the industry, but they were also just not usable the current app. In the end I guided him through the process, and explained the whole picture of how to implement that feature. Which barely took a few lines of code. As a freelancer I'm also really excited with the REAL stuff that is coming with AI like Cursor, this really helps ship a app really fast. Saves a lot of work.

Hahahahaha!

I enjoyed this video so much. Educative top notch 👌

Keep going

data_science != happy life