In [1]:
Tue 12 Oct 2021 01:04:13 PM EDT

update on machine learning blurbs

felt like an update was needed because i'm sure you're all starving for more machine learning morsels.

the farther i get into the first chapter of this book i realize it's a lot of information you would probably cover to give you scope on the rest of the information you'll cover whence doing the machine learning activities.

overview things like being aware of your data - insufficient data for training and/or that data being misrepresentative of what your machine learning algorithm will encounter in production, overfitting and underfitting and tuning hyperparameters and model selection and blah blah blah...

through reading and learning some of this i've come to understand that without some sort of examples (like how i have previously mentioned when the author simply name drops a half page list of algorithm names to describe how to categorize algorithms ) or further context, all of this is simply attempting to rewrite parts of the book without plagiarizing.

which is really really hard and also not really an effective tool for my learning and super extremely duper not effective for your entertainment. not that i'm concerned if you're entertained or not. but it would be nice right?

basically what i am saying is that i thought this would be a good idea but i think i can come up with a better way.

so i'll just drop an update of where my head is at in learning this material from this book.

so far the book starts with some overview concepts that might be helpful along the way when learning how to do the (ML) machine learning things. but because i haven't done any machine learning things none of these concepts really mean anything to me without the practice or application of said ML things.

additionally what i have noticed is that i started using the word algorithms and even attempted to explain it but there is no example anywhere on my site, github, or wherever else, that i put an algorithm into practice. or use computer science-y fancy shmancy algorithms to do anything.

so i might start writing about that too

so if it sounds like i'm stuck on the first chapter of this book because there is all this learning and no doing, that's exactly what's happening.

the second chapter of the book states that it is an end to end machine learning project, as if i was hired as a data scientist and was given their first project. i am excited for this as it seems like it's more directed and there is more doing.

at the end of the first chapter, there are questions to test your knowledge of what was gone over in the first chapter so i might just skip the reading and answer the questions by paging through the chapter and specifically looking for them because we (we being those esteemed folks in a large amount of college debt) all know this is exactly how you get a college degree.

i've probably said this before but this concept overview type teaching might work for some people but i come from the house of doing and doing it wrong to know what the correct way is and then coming back to visit the whys and hows. sometimes though, through doing you're able to discover the whys and hows on your own.

if you're not my mom and you're a programmer thinking about getting the hands on machine learning book, i would say to avoid being discouraged - start on chapter two.

that's it. no valuable insights today about how robots do their homework, just a personal anecdote about me getting mildly frustrated with following someone else's roadmap to learning.

cheers, and thanks for reading.