7:15 AI vs ML vs DL vs DS 17:00 Machine Learning 17:49 Types of ML 19:00 Supervised ML 21:50 Regression intro (brief) 24:04 Classification (brief) 25:35 Unsupervised ML 26:08 Clustering (brief) 28:49 Dimentionality Reduction 33:10 First Algo - Linear Regression
After three year I am Feel like a real school student I am missing those days 😃 i have pen in my hand my notebook and a teacher like you and note and all this thing you teach like in we note and learn in school thankyou so much for giving us enviornment. 🤞
00:14 Introduction to machine learning algorithms for data science 02:48 Introduction to machine learning algorithms and interview preparation 07:32 Artificial intelligence applications can perform tasks without human intervention. 09:57 Applications of AI in providing personalized recommendations and services 14:24 Deep learning aims to mimic the human brain and has enabled solving complex use cases. 16:26 Machine learning involves supervised and unsupervised algorithms. 20:27 Understanding independent and dependent features in supervised machine learning 22:28 Dependent feature determines problem type 26:33 Clustering helps in grouping similar data for customer segmentation 28:27 Introduction to Machine Learning Algorithms and Data Science 32:28 Introduction to linear regression for modeling data 34:56 Introduction to linear regression and hypothesis in machine learning 38:49 Understanding the concepts of theta 0 and theta 1 in machine learning. 40:53 Understanding the concept of slope in linear regression 44:47 Cost function for machine learning algorithms 46:37 Cost function helps in finding the best fit line by minimizing the distance. 50:53 Minimizing the squared error function with parameters theta 0 and theta 1 53:00 Understanding the hypothesis with intercept at origin 56:47 Theta 1 values impact cost function 58:38 Calculation of cost function using gradient descent 1:02:57 Explaining the concept of gradient descent in machine learning. 1:04:59 Convergence algorithm for reaching global minima in gradient descent. 1:09:02 Positive slope indicates weight update in convergence algorithm 1:10:53 Learning rate is crucial for reaching the global minima. 1:14:51 Cost function prevents local minima in linear regression 1:16:50 Understanding the derivative of theta j 1:21:22 Convergence algorithm for gradient descent 1:24:14 Understanding r square and adjusted r square for model performance evaluation 1:29:26 Explanation on why the mean of a particular value distance will be higher 1:31:36 Adding uncorrelated features can increase R-squared value. 1:35:41 Number of samples, predictors, and their impact on R-squared and adjusted R-squared. 1:37:33 R-squared value decreases as the number of predictors increases 1:42:34 Discussion on research papers and future topics in the course. 1:44:39 Encouragement to share and join community session
Please also cover other important topics like LDA,PCA, T-sne etc. A lot of good companies ask these questions and test the in-depth understanding of the candidates. @Krish Naik.
Sir thank you for this great job. Please sir i have a question? Are we to always assume theta one while finding the gradient descent curve? Thank you so much. I love the way you split everything for us. You are so much.
Why 1/2m means dy/dx which means the small changes in j(theta1)/theta1 with respect to theta 1 so that's why 1/2m to reduce to minimum distance between actual minus predict if we do 1/m it will give us big difference so thats why we should not do average instead of average we multiply it with 2 to get slight minimal difference of SSE if divide it with m which means it will get avarage it will just affect the minimal error so that why to reduce the error we divide ➗ 2m which will give 0.0like this so all data points now closest to best fit line this is called cost function
Hi Krish, If there any defination related to that before the explanation it will quite great. What is unsupervised learning? before giving an explanation if there is an defination then we can understand what ur trying to explain. Now it is bit confusing.
Hi sir thanks for sharing your amazing knowledge, i have a doubt at 43.00 min, here wchich difference we will take to get perfect line. is it difference between predictied points or predicted point to existance data set values?
1:00:00 during the calculation of cost function you used 1/2m, which is a normal sum calculation. but you said previously that for simplifying differentiation operation we are using 1/2m, so isn't 1/2 is unnecessary . Just 1/m would be sufficient. Also in standard books, the cost is calculated using (1/total data points) or in your case 1/m. Please clear this.
1/2 * x^2 and then @krish Naik u cancelled denominator 2 with that power. I did not understand this. if x = 4 then it will be 1/2 * x^2 = 1/2 * 4 * 4 but in your case it is just 4 Please explain this
sir how data is prepared.. which we use for machine learning ..like hand writting images dataste..,iris_dataset .. sir can we create the own data like that datsets ..???? can we create datasets like that using python..??
Sir, I have a question. As the R square adjusted always decreases with the increase in predictors, how can we be sure that it is not ignoring features that are actually correlated?
If your r2 adjusted decreased when you add a column it means that the added features is not correlated with the output feature (which needs to be predicted i.e in the above case - house) and can give inappropriate results but If your r2 adjusted increases or is similar to previous one when you add a column it means that the added features is somehow correlated with the output feature and is considered as a better model. Hope the explanation helps..
@@sudhanshuedu wow we're so excited....some. Ofus can't afford one neuron but y'all helping us and m so blessed tht I found ineuron I don't have system yet so m just practicing python On my android phone....hope nearby april I get my own pc or laptop and you guy will launch some fsds job guarantee program nearby summers 😭😭❤️❤️❤️❤️ Hatsoff to you guys 👍🌹
@@sudhanshuedu sudhashu sir thank you for oneneuron platform.. started data science masters .again thank you for an amazing concept of affordable education.
Really amazing tutorial In youtube sufficient content is available for entry level in data science Here some good teacher like Krish But we have maintaine consistency and always be excited to learn new things, skills, and anything And enjoy while learning things because you will be always be grow in professional career if you adopt learning as a hobby
Superb sir i thoroughly enjoyed of learning machine learning. Your energy and practical way of teaching makes more eagerness towards learning data science
Hello sir, I am already an enrolled student of One neuron and also started watching your series on ML on youtube, I have one query regarding Gradient descent, though I have understood the maths behind it I really want to know Do I need to understand its implementation in Python too, in Depth or just the understanding of ALGORITHM WOULD SUFFICE IN INTERVIEWS AND DURING JOBS?
Mr. Krish Naik. It was wonderful listening the theory. Really enjoyed. I belong to mechanical background and a beginner in AI. I have one doubt sir. In finding "Rsquared" you said "ycap" is the difference between the "actual" and the "predicted" point. Then "ycap" is nothing but "htheta x - y" right?. But you said as ycap is nothing but htheta x. It is confusing me. please clear this tomorrow.
Sir I don't know why people use optimisation algorithms in regression to find coefficient whereas already it is proven that coefficient=inv(x'x)%(x'y) which is proven by likelihood function method Just enter value of x and y and and you will get estimate of coefficient no need to do iterative optimization
This method normally works when the dimension of X is within a certain limit. Say if size of X is 10^4 or greater then finding inverse becomes expensive, and the machine fails to do so. But if we use other algo like gradient descent then we can find optimal x even for large value of X.
Hi Sir, Where we can find materials . As have try to open the link provided in Description .Its showing page not avaliable. Plz help with the material.
krish sir at 1:13:44 why are we taking the slope of the local minima ? obv it will be 0 even for global minima shouldn't we check for the that particular green point ? can anyone clarify this doubt ????
SIr, how will the computer know if it has reached then global minima? It will have to have the value or some kind of comparision right? in order to check if it has reached, and if it does, what is the difference between this and finding the smallest value in an array? they both looks the same to me, please clarify sir.
Hello Krish, how are you ? I am new to ML line now and would like to start my journey in ML. I am an experienced employee having several years of IT experience. Do you think if I start learning ML now, it will be beneficial to me to crack jobs ? Because many colleges and institutions are giving training on AIML now a days. Please suggest !!
I am not able to proceed with the community session, maybe because my phone number is a french number. SO can anyone please help me in that? Also if anyone can send me the pdf I will be very obliged
7:15 AI vs ML vs DL vs DS
17:00 Machine Learning
17:49 Types of ML
19:00 Supervised ML
21:50 Regression intro (brief)
24:04 Classification (brief)
25:35 Unsupervised ML
26:08 Clustering (brief)
28:49 Dimentionality Reduction
33:10 First Algo - Linear Regression
Tnq krish.. You are the krish without cape and mask in edtech sector.. Tnq
R sq and adjusted r sq too
l
k
l
k
k
0
After three year I am Feel like a real school student I am missing those days 😃 i have pen in my hand my notebook and a teacher like you and note and all this thing you teach like in we note and learn in school thankyou so much for giving us enviornment. 🤞
I persuing my data science degree but i never learned like learned from you 🖤
cant open your comunity session , page has crashed . help
Great lecture 😊😊
KRISH is a very good teacher
00:14 Introduction to machine learning algorithms for data science
02:48 Introduction to machine learning algorithms and interview preparation
07:32 Artificial intelligence applications can perform tasks without human intervention.
09:57 Applications of AI in providing personalized recommendations and services
14:24 Deep learning aims to mimic the human brain and has enabled solving complex use cases.
16:26 Machine learning involves supervised and unsupervised algorithms.
20:27 Understanding independent and dependent features in supervised machine learning
22:28 Dependent feature determines problem type
26:33 Clustering helps in grouping similar data for customer segmentation
28:27 Introduction to Machine Learning Algorithms and Data Science
32:28 Introduction to linear regression for modeling data
34:56 Introduction to linear regression and hypothesis in machine learning
38:49 Understanding the concepts of theta 0 and theta 1 in machine learning.
40:53 Understanding the concept of slope in linear regression
44:47 Cost function for machine learning algorithms
46:37 Cost function helps in finding the best fit line by minimizing the distance.
50:53 Minimizing the squared error function with parameters theta 0 and theta 1
53:00 Understanding the hypothesis with intercept at origin
56:47 Theta 1 values impact cost function
58:38 Calculation of cost function using gradient descent
1:02:57 Explaining the concept of gradient descent in machine learning.
1:04:59 Convergence algorithm for reaching global minima in gradient descent.
1:09:02 Positive slope indicates weight update in convergence algorithm
1:10:53 Learning rate is crucial for reaching the global minima.
1:14:51 Cost function prevents local minima in linear regression
1:16:50 Understanding the derivative of theta j
1:21:22 Convergence algorithm for gradient descent
1:24:14 Understanding r square and adjusted r square for model performance evaluation
1:29:26 Explanation on why the mean of a particular value distance will be higher
1:31:36 Adding uncorrelated features can increase R-squared value.
1:35:41 Number of samples, predictors, and their impact on R-squared and adjusted R-squared.
1:37:33 R-squared value decreases as the number of predictors increases
1:42:34 Discussion on research papers and future topics in the course.
1:44:39 Encouragement to share and join community session
As always best lecture
sir please do community session on deep learning in future too.
Please also cover other important topics like LDA,PCA, T-sne etc. A lot of good companies ask these questions and test the in-depth understanding of the candidates.
@Krish Naik.
Thank you so much sir making these usefull vedios.awesome explanation.
1:33:33 that moment of realisation when life starts to make sense.
love u sir
Sir thank you for this great job. Please sir i have a question? Are we to always assume theta one while finding the gradient descent curve? Thank you so much. I love the way you split everything for us. You are so much.
Why 1/2m means dy/dx which means the small changes in j(theta1)/theta1 with respect to theta 1 so that's why 1/2m to reduce to minimum distance between actual minus predict if we do 1/m it will give us big difference so thats why we should not do average instead of average we multiply it with 2 to get slight minimal difference of SSE if divide it with m which means it will get avarage it will just affect the minimal error so that why to reduce the error we divide ➗ 2m which will give 0.0like this so all data points now closest to best fit line this is called cost function
Thanks @krish
Superb explanation Thank you so much
46:34 .. cost Function is wrong Sir , The right one is mention here
Cost Function(J)=\frac{1}{n}\sum_{i=0}^{n}(h_{\Theta} (x^{i})-y^{i})^{2}
yes
My guru🙏
At 1:23:13, when he says Oh My god after writing derivation 🙈
Hi Krish, If there any defination related to that before the explanation it will quite great. What is unsupervised learning? before giving an explanation if there is an defination then we can understand what ur trying to explain. Now it is bit confusing.
Hi sir thanks for sharing your amazing knowledge, i have a doubt at 43.00 min, here wchich difference we will take to get perfect line. is it difference between predictied points or predicted point to existance data set values?
Please add this to OneNeuron OTT and named it as "7 day ML - Krish Naik", "7 day Statistics - Krish Naik"..etc. Thanks.
Awesome session!
1:00:00 during the calculation of cost function you used 1/2m, which is a normal sum calculation.
but you said previously that for simplifying differentiation operation we are using 1/2m, so isn't 1/2 is unnecessary . Just 1/m would be sufficient.
Also in standard books, the cost is calculated using (1/total data points) or in your case 1/m.
Please clear this.
The two in 1/2m is to cancel out the differential if X square
we need this hand written material. The above link is not working
Hello Sir, I cant find slides of this lecture can you please share the same?
Even I am looking for the same. I have emailed them last week but didn't get any revery yet. Please let me know if you get the files.
How do i access the materials, the link is broken please let me know
Sir, could you share that slides in the comment or description section please!
Sir, can you confirm in R and R square in short
1/2 * x^2 and then @krish Naik u cancelled denominator 2 with that power. I did not understand this. if x = 4 then it will be 1/2 * x^2 = 1/2 * 4 * 4 but in your case it is just 4
Please explain this
sir where is note
which text eidtor r u using sir ?
What does this symbol := means
what is the derivatives of x of n?I mean what is derivatives?
sir how data is prepared.. which we use for machine learning ..like hand writting images dataste..,iris_dataset ..
sir can we create the own data like that datsets ..????
can we create datasets like that using python..??
can anyone share the notes, i couldn't find it on the provided link?
hello sir, from where can we get notes of the video.
Sir, I have a question.
As the R square adjusted always decreases with the increase in predictors, how can we be sure that it is not ignoring features that are actually correlated?
If your r2 adjusted decreased when you add a column it means that the added features is not correlated with the output feature (which needs to be predicted i.e in the above case - house) and can give inappropriate results
but If your r2 adjusted increases or is similar to previous one when you add a column it means that the added features is somehow correlated with the output feature and is considered as a better model.
Hope the explanation helps..
@@kunjjani1683 Thank you for the explanation! Now I can understand better.
Can someone please share the link to the slides he is referring in the tutorial.
hey krish, you can share your blue note book PDF
For us you're our Andrew NG
Yea🎉
Founder of deep learning 😊
True
Who is Andrew ng
yes
Unable to find the notes in community session. If any one has pls share gdrive link.🙏🙏🙏🙏🙏🙏
Even I am looking for the same. I have emailed them last week but didn't get any revery yet. Please let me know if you get the files.
Bring DL, Maths for DS and NLP 7 Days session on each of these topics.
Notes available h isske??
I can't find the resources/ notes in community session?
Thanks Krish,, your blue book inspired me to start writing notes in my book.
What's blue book?
7:50
Sir will we be doing the coding portion of the algorithms as well in python in these 7 days time?
yes i will be doing an implementation part dont worry .
@@sudhanshuedu wow we're so excited....some. Ofus can't afford one neuron but y'all helping us and m so blessed tht I found ineuron
I don't have system yet so m just practicing python On my android phone....hope nearby april I get my own pc or laptop and you guy will launch some fsds job guarantee program nearby summers 😭😭❤️❤️❤️❤️
Hatsoff to you guys 👍🌹
@@sudhanshuedu where will you Upload Coding?
@@sudhanshuedu sudhashu sir thank you for oneneuron platform.. started data science masters .again thank you for an amazing concept of affordable education.
@sudhanshu kumar thank you sir
Really amazing tutorial
In youtube sufficient content is available for entry level in data science
Here some good teacher like Krish
But we have maintaine consistency and always be excited to learn new things, skills, and anything
And enjoy while learning things because you will be always be grow in professional career if you adopt learning as a hobby
you can also checkout Campus-X, he has very good in depth lectures.
Great course..Krish can you open this for downloading??
None of the numbers in the description have responded.
Kindly let me know How can Ik more about the course?
KNN should come in Supervised Ml
Can anyone help me to tell me about that in the way of learning data science is there any domain need to be learn like nlp, cv etc..????
Where is it's concept implementation, without implementation how we can relate... 😢
Very difficult to understand math and equations for Linear Regression
can you share notes?
Thanks a lot for this...however, I am not able to find the pdf notes for these sessions. Can anyone help?
You are very good teacher because I'm making understand the algorithm of ml😁🙏
I visited the mentioned website. But i am in trouble to find the notes. Anybody here to find this ?
By the way, i am very much enjoying this tutorial.
same doubt
I m unable to find the notes pls upload the material
Sir please do 1.30 theoretical part and rest 15 minutes on implementation
Superb sir i thoroughly enjoyed of learning machine learning. Your energy and practical way of teaching makes more eagerness towards learning data science
Hello sir,
I am already an enrolled student of One neuron and also started watching your series on ML on youtube, I have one query regarding Gradient descent, though I have understood the maths behind it I really want to know Do I need to understand its implementation in Python too, in Depth or just the understanding of ALGORITHM WOULD SUFFICE IN INTERVIEWS AND DURING JOBS?
After the video please comment the syllabus of the day or at least name of algorithms covered. Thankyou!
Check the thumbnail😀😀😀
Mr. Krish Naik. It was wonderful listening the theory. Really enjoyed. I belong to mechanical background and a beginner in AI.
I have one doubt sir. In finding "Rsquared" you said "ycap" is the difference between the "actual" and the "predicted" point. Then "ycap" is nothing but "htheta x - y" right?. But you said as ycap is nothing but htheta x. It is confusing me. please clear this tomorrow.
@@ArsalanKhan-qyy Sorry I am not clear
its clear but am confused because to the speed of your lecturing.
Please do Coding of each Algorithm
Very nice recap of all stuff in one place. Tumba tumba Dhanyavadagalu😊🙏
if this could with the practical on Jupyter notebook, that would be amazing too.
but still, Thank you sir, these efforts are amazing.
can u please suggest the name of the digital pad and pen u use?
Sir I don't know why people use optimisation algorithms in regression to find coefficient whereas already it is proven that coefficient=inv(x'x)%(x'y) which is proven by likelihood function method
Just enter value of x and y and and you will get estimate of coefficient no need to do iterative optimization
matrix inverse is not always defined and also it is very costly operation
This method normally works when the dimension of X is within a certain limit. Say if size of X is 10^4 or greater then finding inverse becomes expensive, and the machine fails to do so. But if we use other algo like gradient descent then we can find optimal x even for large value of X.
Does anyone have notes for this video , if yes please share
Krish, which textbook would you recommend alongwith these videos to understand the matter in detail.
atleast for paid members remove advertisements
the notepad available for download??
hlo sir this is raju naik from hyderabad
Does IIT provides Btech in data science 🙏🙏🙏
Check IIT Madras online degree program, They have launched BSc in Data Science and Programming
Sir, a lot of respect to you, Excellent session :)
could anybody please help me understand why we did 1/m and not 1/2m for convergence equation?
It got cancelled by the two in the differential
Hi Sir, Where we can find materials . As have try to open the link provided in Description .Its showing page not avaliable. Plz help with the material.
Excellent 1st session. Excited to do the 2nd part tomorrow. I am trying to revise my ML knowledge to get back to competing in Kaggle.
Hi Joseph, How exactly, R u using Kaggle, Can u tell me, I was thinking to participate inHackathon, but never used Kaggle before.
Krish,You are amazing!
krish sir at 1:13:44 why are we taking the slope of the local minima ? obv it will be 0 even for global minima
shouldn't we check for the that particular green point ?
can anyone clarify this doubt ????
SIr, how will the computer know if it has reached then global minima? It will have to have the value or some kind of comparision right? in order to check if it has reached, and if it does, what is the difference between this and finding the smallest value in an array? they both looks the same to me, please clarify sir.
It is really great session. It took me around 4 hours to complete.
Creating my Orange book XD
great efforts(🙏)..... lack of presentation, lengthy & confusing explanation
Hello Krish, how are you ? I am new to ML line now and would like to start my journey in ML. I am an experienced employee having several years of IT experience. Do you think if I start learning ML now, it will be beneficial to me to crack jobs ? Because many colleges and institutions are giving training on AIML now a days.
Please suggest !!
Material link
I'm given to a lots of thanks from bottom my 💓💓💓💖💖
Sir Hindi me banaye
I am not able to proceed with the community session, maybe because my phone number is a french number. SO can anyone please help me in that? Also if anyone can send me the pdf I will be very obliged
From where I can access your notes ?
Please help
Because writing notes is very tedious....
This virtual notebook is amazing 💪🏻🔥 ppl like me will start writing notes now 😂
Name of this notebook?
Sir video banaiye
why you took sqare ???. we can take mod of that function so that it will give positive value???
Hi sir
For linear regression cost function only divided by 2 what if the derivative has 3 in numerater then how 3 is divided by 2.
Hi Krish.. Not able to find notes through link you have provided. Can you please update the same there.
25:57