How do I approach learning Data Science/ML the 'rightest' way? [closed] The Next CEO of Stack Overflow2019 Community Moderator ElectionWhat skills do I need to become a data scientist? And how to show them?data science / machine learning resources?Using machine learning specifically for feature analysis, not predictionsWhat are the 'hottest' future areas of Machine Learning and Data Science?Any guidance for new beginners interested in data scienceTopics to cover for software developer interested in data analyticsIntro to Machine LearningMachine Learning to predict risk of itemsUsing packages such as sklearn vs building ML algorithm from scratchTools and Techniques for Analyzing German Automotive Discussion Forum PostsWhich Kind of Machine Learning should I use for an Optimization Problem?
Planeswalker Ability and Death Timing
How to show a landlord what we have in savings?
Which acid/base does a strong base/acid react when added to a buffer solution?
My boss doesn't want me to have a side project
Is it reasonable to ask other researchers to send me their previous grant applications?
MT "will strike" & LXX "will watch carefully" (Gen 3:15)?
Upgrading From a 9 Speed Sora Derailleur?
Can a PhD from a non-TU9 German university become a professor in a TU9 university?
Mathematica command that allows it to read my intentions
What is the difference between 'contrib' and 'non-free' packages repositories?
How to pronounce fünf in 45
Can Sri Krishna be called 'a person'?
Does Germany produce more waste than the US?
Man transported from Alternate World into ours by a Neutrino Detector
Car headlights in a world without electricity
Is the offspring between a demon and a celestial possible? If so what is it called and is it in a book somewhere?
Creating a script with console commands
Raspberry pi 3 B with Ubuntu 18.04 server arm64: what pi version
Is there a rule of thumb for determining the amount one should accept for a settlement offer?
Is it possible to create a QR code using text?
Is it possible to make a 9x9 table fit within the default margins?
Would a grinding machine be a simple and workable propulsion system for an interplanetary spacecraft?
That's an odd coin - I wonder why
Incomplete cube
How do I approach learning Data Science/ML the 'rightest' way? [closed]
The Next CEO of Stack Overflow2019 Community Moderator ElectionWhat skills do I need to become a data scientist? And how to show them?data science / machine learning resources?Using machine learning specifically for feature analysis, not predictionsWhat are the 'hottest' future areas of Machine Learning and Data Science?Any guidance for new beginners interested in data scienceTopics to cover for software developer interested in data analyticsIntro to Machine LearningMachine Learning to predict risk of itemsUsing packages such as sklearn vs building ML algorithm from scratchTools and Techniques for Analyzing German Automotive Discussion Forum PostsWhich Kind of Machine Learning should I use for an Optimization Problem?
$begingroup$
First of all, I am not sure if this is the right place to post this so please do let me know if it isn't and tell me where it should be. I really just don't know where to go with this question.
Some backstory: I am going through my 2nd semester of my 2nd year of Software Engineering. I love math, I love computers, I love data and I love image processing so the most logical place for me to be is Data Science because I seem to be gravitating towards ML/CV a lot.
Thing is, I want to start learning Data Science/ML but I don't know where to start. Everyone keeps recommending online courses such as Andrew's coursera course but I have become somewhat skeptical of most online courses because they seem to love simplifying info too much that I feel like it just passes the threshold of it being useful info.
My assumption is: learning Data Science/ML/CV requires a very strong and rigorous foundation. I shouldn't start learning it with high level tools and little understanding of what makes it what it is. This is my problem. I don't know where to begin learning that way.
I searched for books and the two books I found people talking about were:
- Pattern Recognition and Machine Learning, by Christopher M. Bishop. This book seemed to be very rigorous with a strong focus on the mathematics and intuition of things (looking at its index). Not much skepticism towards this one.
- Hands-On Machine Learning with Scikit & TensorFlow, by Aurelien Geron. This one seems to be extremely popular and almost unanimously praised/liked/recommended by people but just looking at its name I felt a bit skeptical that it might have less focus on building strong foundation/intuition and more focus on using high-level tools.
I am open for other suggestions if you think some 3rd book/resource is more adequate for my need. I'd also greatly appreciate it if you briefly explain why you think it is more adequate.
Also it's worth noting that I have no problem with high-level tools and I know that I will eventually have to use them. I just feel like I'd be significantly better at what I want to do and valuable if I start from there.
machine-learning learning
$endgroup$
closed as primarily opinion-based by Siong Thye Goh, Simon Larsson, Mark.F, oW_, Sean Owen♦ Mar 26 at 3:08
Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise. If this question can be reworded to fit the rules in the help center, please edit the question.
add a comment |
$begingroup$
First of all, I am not sure if this is the right place to post this so please do let me know if it isn't and tell me where it should be. I really just don't know where to go with this question.
Some backstory: I am going through my 2nd semester of my 2nd year of Software Engineering. I love math, I love computers, I love data and I love image processing so the most logical place for me to be is Data Science because I seem to be gravitating towards ML/CV a lot.
Thing is, I want to start learning Data Science/ML but I don't know where to start. Everyone keeps recommending online courses such as Andrew's coursera course but I have become somewhat skeptical of most online courses because they seem to love simplifying info too much that I feel like it just passes the threshold of it being useful info.
My assumption is: learning Data Science/ML/CV requires a very strong and rigorous foundation. I shouldn't start learning it with high level tools and little understanding of what makes it what it is. This is my problem. I don't know where to begin learning that way.
I searched for books and the two books I found people talking about were:
- Pattern Recognition and Machine Learning, by Christopher M. Bishop. This book seemed to be very rigorous with a strong focus on the mathematics and intuition of things (looking at its index). Not much skepticism towards this one.
- Hands-On Machine Learning with Scikit & TensorFlow, by Aurelien Geron. This one seems to be extremely popular and almost unanimously praised/liked/recommended by people but just looking at its name I felt a bit skeptical that it might have less focus on building strong foundation/intuition and more focus on using high-level tools.
I am open for other suggestions if you think some 3rd book/resource is more adequate for my need. I'd also greatly appreciate it if you briefly explain why you think it is more adequate.
Also it's worth noting that I have no problem with high-level tools and I know that I will eventually have to use them. I just feel like I'd be significantly better at what I want to do and valuable if I start from there.
machine-learning learning
$endgroup$
closed as primarily opinion-based by Siong Thye Goh, Simon Larsson, Mark.F, oW_, Sean Owen♦ Mar 26 at 3:08
Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise. If this question can be reworded to fit the rules in the help center, please edit the question.
add a comment |
$begingroup$
First of all, I am not sure if this is the right place to post this so please do let me know if it isn't and tell me where it should be. I really just don't know where to go with this question.
Some backstory: I am going through my 2nd semester of my 2nd year of Software Engineering. I love math, I love computers, I love data and I love image processing so the most logical place for me to be is Data Science because I seem to be gravitating towards ML/CV a lot.
Thing is, I want to start learning Data Science/ML but I don't know where to start. Everyone keeps recommending online courses such as Andrew's coursera course but I have become somewhat skeptical of most online courses because they seem to love simplifying info too much that I feel like it just passes the threshold of it being useful info.
My assumption is: learning Data Science/ML/CV requires a very strong and rigorous foundation. I shouldn't start learning it with high level tools and little understanding of what makes it what it is. This is my problem. I don't know where to begin learning that way.
I searched for books and the two books I found people talking about were:
- Pattern Recognition and Machine Learning, by Christopher M. Bishop. This book seemed to be very rigorous with a strong focus on the mathematics and intuition of things (looking at its index). Not much skepticism towards this one.
- Hands-On Machine Learning with Scikit & TensorFlow, by Aurelien Geron. This one seems to be extremely popular and almost unanimously praised/liked/recommended by people but just looking at its name I felt a bit skeptical that it might have less focus on building strong foundation/intuition and more focus on using high-level tools.
I am open for other suggestions if you think some 3rd book/resource is more adequate for my need. I'd also greatly appreciate it if you briefly explain why you think it is more adequate.
Also it's worth noting that I have no problem with high-level tools and I know that I will eventually have to use them. I just feel like I'd be significantly better at what I want to do and valuable if I start from there.
machine-learning learning
$endgroup$
First of all, I am not sure if this is the right place to post this so please do let me know if it isn't and tell me where it should be. I really just don't know where to go with this question.
Some backstory: I am going through my 2nd semester of my 2nd year of Software Engineering. I love math, I love computers, I love data and I love image processing so the most logical place for me to be is Data Science because I seem to be gravitating towards ML/CV a lot.
Thing is, I want to start learning Data Science/ML but I don't know where to start. Everyone keeps recommending online courses such as Andrew's coursera course but I have become somewhat skeptical of most online courses because they seem to love simplifying info too much that I feel like it just passes the threshold of it being useful info.
My assumption is: learning Data Science/ML/CV requires a very strong and rigorous foundation. I shouldn't start learning it with high level tools and little understanding of what makes it what it is. This is my problem. I don't know where to begin learning that way.
I searched for books and the two books I found people talking about were:
- Pattern Recognition and Machine Learning, by Christopher M. Bishop. This book seemed to be very rigorous with a strong focus on the mathematics and intuition of things (looking at its index). Not much skepticism towards this one.
- Hands-On Machine Learning with Scikit & TensorFlow, by Aurelien Geron. This one seems to be extremely popular and almost unanimously praised/liked/recommended by people but just looking at its name I felt a bit skeptical that it might have less focus on building strong foundation/intuition and more focus on using high-level tools.
I am open for other suggestions if you think some 3rd book/resource is more adequate for my need. I'd also greatly appreciate it if you briefly explain why you think it is more adequate.
Also it's worth noting that I have no problem with high-level tools and I know that I will eventually have to use them. I just feel like I'd be significantly better at what I want to do and valuable if I start from there.
machine-learning learning
machine-learning learning
asked Mar 25 at 12:26
Eyad H.Eyad H.
1062
1062
closed as primarily opinion-based by Siong Thye Goh, Simon Larsson, Mark.F, oW_, Sean Owen♦ Mar 26 at 3:08
Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise. If this question can be reworded to fit the rules in the help center, please edit the question.
closed as primarily opinion-based by Siong Thye Goh, Simon Larsson, Mark.F, oW_, Sean Owen♦ Mar 26 at 3:08
Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise. If this question can be reworded to fit the rules in the help center, please edit the question.
add a comment |
add a comment |
2 Answers
2
active
oldest
votes
$begingroup$
Welcome to the community!
I normally don't answer questions like this, they're pretty broad and there are many others just like it. It seems like you're not actually asking how to get started, you already seem set on reading some books (which is great!), so let me answer that question, rather than a very broad 'how to get started'.
First of all, I'd like to challenge your assumption about learning from the bottom up, and offer an alternative opinion (yes, this part is just an opinion). I'd instead encourage you to get your hands on the high level tools and start using them right away. If everything we ever did required us to forego abstractions and learn everything from the beginning, we'd still be reinventing the wheel over and over. Start experimenting with the existing tools and supplement that with your reading. You'll develop an excellent understanding this way.
So to answer your question... I'd recommend the second book you quoted: Hands-On ML with Scikit Learn and Tensorflow. This book will teach you good habits right from the beginning, give you a good understanding, and get you using those tools right away.
Best of luck.
$endgroup$
add a comment |
$begingroup$
I tryed to give an answer to similar question in What skills do I need to become a data scientist? And how to show them?, there are a lot of online courses and they are pretty usefull but if you feel like they are oversimplified you can try other courses from Stanford and Mit for example, like:
- Stanfords Machine Learning Class
- Introduction to convolutional neural networks for visual recognition
Also, Data Science performs better when you have basic knowledge of the subject area, Digital Image Processing from Rich Radke will give you the basics to understand CV algorithms, he also has a course in Computer Vision for Visual Effects that is marvellours.
Udacity and other online courses come from universities. Is important to note that Data Science is academia is still getting confy and until there the courses will discuss a lot about that is really relevant to teach.
Don't get me wrong, books are lovely but if you look for rigorous books such as Bishop's you might be a little outdated, while if you only use books such as Data Science from Scracth and Data Smart your knowledge might still be a bit shallow.
I suggest you try simplied crash courses such as Google's Machine Learning Crash Course but after that don't fall into the illusion you know everything.
Most of theses shallow courses will get you ready for basic tasks and you can improve your knowledge gradually by making Kaggle's competitions and more deep courses later. If you don't get your hands on doing you might feel a bit disapointed and lose motivation.
It is good to learn everything from the beggining but that is after you get knowledge of high level stuff. YOU SHOULD NEVER DO SOMETHING THAT ALREADY EXISTS FROM ZERO WHEN YOU ARE WORKING. Don't be offended your code will surelly be bad and unefficient at first, so unless popular libraries are REALLY lacking in functionality you need, use them.
Note: You should try to get your hands on with OpenCV and Algorithms such as Random Forests and such. Don't fall for Deep Learning yet, it is easy for you bud really expensive for computers and organizations since they usually take long to train and if you're a good programmer you know that most things don't work in the first time around.
$endgroup$
add a comment |
2 Answers
2
active
oldest
votes
2 Answers
2
active
oldest
votes
active
oldest
votes
active
oldest
votes
$begingroup$
Welcome to the community!
I normally don't answer questions like this, they're pretty broad and there are many others just like it. It seems like you're not actually asking how to get started, you already seem set on reading some books (which is great!), so let me answer that question, rather than a very broad 'how to get started'.
First of all, I'd like to challenge your assumption about learning from the bottom up, and offer an alternative opinion (yes, this part is just an opinion). I'd instead encourage you to get your hands on the high level tools and start using them right away. If everything we ever did required us to forego abstractions and learn everything from the beginning, we'd still be reinventing the wheel over and over. Start experimenting with the existing tools and supplement that with your reading. You'll develop an excellent understanding this way.
So to answer your question... I'd recommend the second book you quoted: Hands-On ML with Scikit Learn and Tensorflow. This book will teach you good habits right from the beginning, give you a good understanding, and get you using those tools right away.
Best of luck.
$endgroup$
add a comment |
$begingroup$
Welcome to the community!
I normally don't answer questions like this, they're pretty broad and there are many others just like it. It seems like you're not actually asking how to get started, you already seem set on reading some books (which is great!), so let me answer that question, rather than a very broad 'how to get started'.
First of all, I'd like to challenge your assumption about learning from the bottom up, and offer an alternative opinion (yes, this part is just an opinion). I'd instead encourage you to get your hands on the high level tools and start using them right away. If everything we ever did required us to forego abstractions and learn everything from the beginning, we'd still be reinventing the wheel over and over. Start experimenting with the existing tools and supplement that with your reading. You'll develop an excellent understanding this way.
So to answer your question... I'd recommend the second book you quoted: Hands-On ML with Scikit Learn and Tensorflow. This book will teach you good habits right from the beginning, give you a good understanding, and get you using those tools right away.
Best of luck.
$endgroup$
add a comment |
$begingroup$
Welcome to the community!
I normally don't answer questions like this, they're pretty broad and there are many others just like it. It seems like you're not actually asking how to get started, you already seem set on reading some books (which is great!), so let me answer that question, rather than a very broad 'how to get started'.
First of all, I'd like to challenge your assumption about learning from the bottom up, and offer an alternative opinion (yes, this part is just an opinion). I'd instead encourage you to get your hands on the high level tools and start using them right away. If everything we ever did required us to forego abstractions and learn everything from the beginning, we'd still be reinventing the wheel over and over. Start experimenting with the existing tools and supplement that with your reading. You'll develop an excellent understanding this way.
So to answer your question... I'd recommend the second book you quoted: Hands-On ML with Scikit Learn and Tensorflow. This book will teach you good habits right from the beginning, give you a good understanding, and get you using those tools right away.
Best of luck.
$endgroup$
Welcome to the community!
I normally don't answer questions like this, they're pretty broad and there are many others just like it. It seems like you're not actually asking how to get started, you already seem set on reading some books (which is great!), so let me answer that question, rather than a very broad 'how to get started'.
First of all, I'd like to challenge your assumption about learning from the bottom up, and offer an alternative opinion (yes, this part is just an opinion). I'd instead encourage you to get your hands on the high level tools and start using them right away. If everything we ever did required us to forego abstractions and learn everything from the beginning, we'd still be reinventing the wheel over and over. Start experimenting with the existing tools and supplement that with your reading. You'll develop an excellent understanding this way.
So to answer your question... I'd recommend the second book you quoted: Hands-On ML with Scikit Learn and Tensorflow. This book will teach you good habits right from the beginning, give you a good understanding, and get you using those tools right away.
Best of luck.
answered Mar 25 at 14:28
Dan CarterDan Carter
8251218
8251218
add a comment |
add a comment |
$begingroup$
I tryed to give an answer to similar question in What skills do I need to become a data scientist? And how to show them?, there are a lot of online courses and they are pretty usefull but if you feel like they are oversimplified you can try other courses from Stanford and Mit for example, like:
- Stanfords Machine Learning Class
- Introduction to convolutional neural networks for visual recognition
Also, Data Science performs better when you have basic knowledge of the subject area, Digital Image Processing from Rich Radke will give you the basics to understand CV algorithms, he also has a course in Computer Vision for Visual Effects that is marvellours.
Udacity and other online courses come from universities. Is important to note that Data Science is academia is still getting confy and until there the courses will discuss a lot about that is really relevant to teach.
Don't get me wrong, books are lovely but if you look for rigorous books such as Bishop's you might be a little outdated, while if you only use books such as Data Science from Scracth and Data Smart your knowledge might still be a bit shallow.
I suggest you try simplied crash courses such as Google's Machine Learning Crash Course but after that don't fall into the illusion you know everything.
Most of theses shallow courses will get you ready for basic tasks and you can improve your knowledge gradually by making Kaggle's competitions and more deep courses later. If you don't get your hands on doing you might feel a bit disapointed and lose motivation.
It is good to learn everything from the beggining but that is after you get knowledge of high level stuff. YOU SHOULD NEVER DO SOMETHING THAT ALREADY EXISTS FROM ZERO WHEN YOU ARE WORKING. Don't be offended your code will surelly be bad and unefficient at first, so unless popular libraries are REALLY lacking in functionality you need, use them.
Note: You should try to get your hands on with OpenCV and Algorithms such as Random Forests and such. Don't fall for Deep Learning yet, it is easy for you bud really expensive for computers and organizations since they usually take long to train and if you're a good programmer you know that most things don't work in the first time around.
$endgroup$
add a comment |
$begingroup$
I tryed to give an answer to similar question in What skills do I need to become a data scientist? And how to show them?, there are a lot of online courses and they are pretty usefull but if you feel like they are oversimplified you can try other courses from Stanford and Mit for example, like:
- Stanfords Machine Learning Class
- Introduction to convolutional neural networks for visual recognition
Also, Data Science performs better when you have basic knowledge of the subject area, Digital Image Processing from Rich Radke will give you the basics to understand CV algorithms, he also has a course in Computer Vision for Visual Effects that is marvellours.
Udacity and other online courses come from universities. Is important to note that Data Science is academia is still getting confy and until there the courses will discuss a lot about that is really relevant to teach.
Don't get me wrong, books are lovely but if you look for rigorous books such as Bishop's you might be a little outdated, while if you only use books such as Data Science from Scracth and Data Smart your knowledge might still be a bit shallow.
I suggest you try simplied crash courses such as Google's Machine Learning Crash Course but after that don't fall into the illusion you know everything.
Most of theses shallow courses will get you ready for basic tasks and you can improve your knowledge gradually by making Kaggle's competitions and more deep courses later. If you don't get your hands on doing you might feel a bit disapointed and lose motivation.
It is good to learn everything from the beggining but that is after you get knowledge of high level stuff. YOU SHOULD NEVER DO SOMETHING THAT ALREADY EXISTS FROM ZERO WHEN YOU ARE WORKING. Don't be offended your code will surelly be bad and unefficient at first, so unless popular libraries are REALLY lacking in functionality you need, use them.
Note: You should try to get your hands on with OpenCV and Algorithms such as Random Forests and such. Don't fall for Deep Learning yet, it is easy for you bud really expensive for computers and organizations since they usually take long to train and if you're a good programmer you know that most things don't work in the first time around.
$endgroup$
add a comment |
$begingroup$
I tryed to give an answer to similar question in What skills do I need to become a data scientist? And how to show them?, there are a lot of online courses and they are pretty usefull but if you feel like they are oversimplified you can try other courses from Stanford and Mit for example, like:
- Stanfords Machine Learning Class
- Introduction to convolutional neural networks for visual recognition
Also, Data Science performs better when you have basic knowledge of the subject area, Digital Image Processing from Rich Radke will give you the basics to understand CV algorithms, he also has a course in Computer Vision for Visual Effects that is marvellours.
Udacity and other online courses come from universities. Is important to note that Data Science is academia is still getting confy and until there the courses will discuss a lot about that is really relevant to teach.
Don't get me wrong, books are lovely but if you look for rigorous books such as Bishop's you might be a little outdated, while if you only use books such as Data Science from Scracth and Data Smart your knowledge might still be a bit shallow.
I suggest you try simplied crash courses such as Google's Machine Learning Crash Course but after that don't fall into the illusion you know everything.
Most of theses shallow courses will get you ready for basic tasks and you can improve your knowledge gradually by making Kaggle's competitions and more deep courses later. If you don't get your hands on doing you might feel a bit disapointed and lose motivation.
It is good to learn everything from the beggining but that is after you get knowledge of high level stuff. YOU SHOULD NEVER DO SOMETHING THAT ALREADY EXISTS FROM ZERO WHEN YOU ARE WORKING. Don't be offended your code will surelly be bad and unefficient at first, so unless popular libraries are REALLY lacking in functionality you need, use them.
Note: You should try to get your hands on with OpenCV and Algorithms such as Random Forests and such. Don't fall for Deep Learning yet, it is easy for you bud really expensive for computers and organizations since they usually take long to train and if you're a good programmer you know that most things don't work in the first time around.
$endgroup$
I tryed to give an answer to similar question in What skills do I need to become a data scientist? And how to show them?, there are a lot of online courses and they are pretty usefull but if you feel like they are oversimplified you can try other courses from Stanford and Mit for example, like:
- Stanfords Machine Learning Class
- Introduction to convolutional neural networks for visual recognition
Also, Data Science performs better when you have basic knowledge of the subject area, Digital Image Processing from Rich Radke will give you the basics to understand CV algorithms, he also has a course in Computer Vision for Visual Effects that is marvellours.
Udacity and other online courses come from universities. Is important to note that Data Science is academia is still getting confy and until there the courses will discuss a lot about that is really relevant to teach.
Don't get me wrong, books are lovely but if you look for rigorous books such as Bishop's you might be a little outdated, while if you only use books such as Data Science from Scracth and Data Smart your knowledge might still be a bit shallow.
I suggest you try simplied crash courses such as Google's Machine Learning Crash Course but after that don't fall into the illusion you know everything.
Most of theses shallow courses will get you ready for basic tasks and you can improve your knowledge gradually by making Kaggle's competitions and more deep courses later. If you don't get your hands on doing you might feel a bit disapointed and lose motivation.
It is good to learn everything from the beggining but that is after you get knowledge of high level stuff. YOU SHOULD NEVER DO SOMETHING THAT ALREADY EXISTS FROM ZERO WHEN YOU ARE WORKING. Don't be offended your code will surelly be bad and unefficient at first, so unless popular libraries are REALLY lacking in functionality you need, use them.
Note: You should try to get your hands on with OpenCV and Algorithms such as Random Forests and such. Don't fall for Deep Learning yet, it is easy for you bud really expensive for computers and organizations since they usually take long to train and if you're a good programmer you know that most things don't work in the first time around.
answered Mar 25 at 14:39
Pedro Henrique MonfortePedro Henrique Monforte
1156
1156
add a comment |
add a comment |