Quick extra note for anyone since I was stuck on it for a bit. If you have any columns you want to exclude beforehand (for me, it kept picking up data with string values too and I wanted to exclude those), run this first for those columns before defining x and y: data_df.drop(['Column1', 'Column2', 'etc..'], axis=1, inplace=True) Using "inplace=True" will make it so those columns will stay out of the dataframe because if you define it as "inplace=False" or don't define it at all, those columns you removed will go back into the dataframe anyway. It wasn't used in defining x and y because we need the column PE to return to the dataframe. And thank you so much for this video miss Megha. I'm new to Python and have been struggling with this for an assignment for hours and this really helped me.
Hi Megha...the explanation is neat and clean...right at the point. Very beautifully explained and the concept is clear. Can you please upload more videos on Logistic Regression, KNN, Random forest, Support Vector machines, Decision tree etc?
Hi Megha, you are a mega tutor indeed. You're more than awesome. Without any gaining, this is the best explanation and best perspective I have come across on youtube regarding ML. You're superb ma'am. I await more of your uploads. Thank you!
Hi, your videos are awesome and easy to understand. Can you please upload the logistic regression, random forest, SVM and times series modeling videos with examples.
@@jaydeepraut820 did you try to search in Google for the dataset described, as she showed and explained it at the video timestamp @1:10 and following? As she describes this in the video, it is straightforward to search and find from Google.
I really appreciate your work, however, I will like to point something I noticed out. The scatter plot you created for your result at the end of the analysis where you had the y_test plotted against the y_pred seems inaccurate to me. plt.scatter(y_test, y_pred) is supposed to indicate that your y_test is on the x-axis while your y_pred. is to be plotted on the y-axis. I believe what you should do is showcase the y_test and y_pred on the y-axis while you use a common x-axis for the two on the same plot.
Two different ways of showing the same basic thing. y_pred vs. y_test will form a 45-degree straight line if the data is perfectly predicted, whereas y_test and y_pred both vs. x_test will show the points overlapping each other, if perfectly predicted. However, because this is multiple regression, there isn't a single x variable, so you would have several of the latter plots. These are useful for certain diagnostics. However, a very common first plot is y_test vs. y_pred. Plotting residuals vs. y_test or x_test (one at a time) are also common charts to make.
Really nice explanation. One question, in the final regression equation can we have 0 as coefficient for any independent variable or all variables will be assigned some non 0 values?
It should always have a non zero value because if it is zero it would mean that the specific independent variable is totally useless to predict the dependent variable, which never is the case.
Because you're estimating from a sample, it would be nearly impossible for a variable to have a coefficient of 0, even if there isn't any correlation between the dependent variable and the independent variable (within the context of your model) in the real process. If you ever get a zero coefficient, it's likely due to an error such as including a redundant feature, having perfect (multi)collinearity, including k dummy variables (and an intercept) for k cases, etc. This can be fixed by reducing the feature set (one at a time, of course.)
Hi Megha, thanks for showing steps by steps. I have a question. Instead plotting the result by "Actual" and "Predicted", can we visualize the predicted vs. actual ''y" for each variable"x"? Can you please advise the codes? Thank you.
how to perform regression for the following equation: time= ( pressure)^a + (concentration of fuel)^b + (concentration of oxidizer)^c + exp(d/Temperature). Here, time is dependent variable on pressure, concentration of fuel, concentration of oxidizer and temperature. How to set non-linear model like the above equation
Thanks for the video. I built a model to predict rent across my country. The accuracy score is 43% or so. What can I do to improve it? I can send the script of needed. Thanks.
Could you please make a video on this topic from scratch without using the sklearn library to better understand the mathematics behind it or provide a useful link? Please...
@Chandra Shakhar start with the website she shows in her discussion of sklearn: scikit-learn.org/stable/modules/generated/sklearn.model_selection.train_test_split.html , which leads you to the technical explanation at: scikit-learn.org/stable/modules/cross_validation.html#cross-validation
Hi Megha, thanks for tutorial. what if we have string in datasets (like types can be multiple strings not boolean e.g colors:blue, red, green ,black ) how we will convert it into float format cuz model only understands numbers.
Fantastic tutorial. Question I have is if I wanted to test my model on another dataset how could I do it once I have my coefficients and intercept? Best Wishes Peter
Hi Megha, thanks for this great video, very simple question may be for you, on your predicted values chart, is there a way to plot a straight line across those values?
Brilliantly explained. Can u make video for deployment of model to use with webpage/android or any programming language GUI through API . Also make such a beautifully explained video for ANN also.
thx for video a lot. But what about the prediction of finding most minimum value in according to independents variables. How can we find it? thanks in advance
Literal queen. Been crying for a week over this. I could've just watched this, this is amazing.
After 6 hours of stumbling through StackOverflow and various books, this video made it clear in 20 minutes!
Thank you SO MUCH!!!
Amazing video! This is my first project ever and I hope to continue further in my career of Data Science!
Since 2 days i am trying to understand ML. Finally abhi ache se samajh gaya. Thanks
Quick extra note for anyone since I was stuck on it for a bit.
If you have any columns you want to exclude beforehand (for me, it kept picking up data with string values too and I wanted to exclude those), run this first for those columns before defining x and y:
data_df.drop(['Column1', 'Column2', 'etc..'], axis=1, inplace=True)
Using "inplace=True" will make it so those columns will stay out of the dataframe because if you define it as "inplace=False" or don't define it at all, those columns you removed will go back into the dataframe anyway. It wasn't used in defining x and y because we need the column PE to return to the dataframe.
And thank you so much for this video miss Megha. I'm new to Python and have been struggling with this for an assignment for hours and this really helped me.
Beginning my ML journey. Thank you for the crisp explanation.
This is the best tutorial i have come across ... simple ,easy and beautiful.
Please upload other regression and classification problems.
Simple, Clear, Concise. What else do you want?
Thank you!
one of the best explanations in very simple words... Bravo Miss Megha Narang
Hi Megha...the explanation is neat and clean...right at the point. Very beautifully explained and the concept is clear. Can you please upload more videos on Logistic Regression, KNN, Random forest, Support Vector machines, Decision tree etc?
In wish I can give this video a million likes.... thank you very much.. this video was really helpful.
Thanks sist, you help me to understand about a long code becomes a short code. It's a smart video.
What keyboard are you using? It sounds so soothing.
Thank you for this very informative tutorial! Please keep uploading
Thank you so much for the video!
Thank you for the visualization.... :)
Wow really its very easy to understand, Mam your sequence wise explanation is awesome.
Hi Megha, you are a mega tutor indeed. You're more than awesome. Without any gaining, this is the best explanation and best perspective I have come across on youtube regarding ML. You're superb ma'am. I await more of your uploads.
Thank you!
I only had difficulty in plotting the model, thanks a lot 😃😃👍👍
Hello Megha. Great video. But you did apply the scaler function to standardize the days. Why?
Very meaningful session, great explanation 👍
For a beginner this video is a big help!!
Hi Megha, thank you so much for the video! It helped me a lot in work.
Really appreciate! hope you keep making that
Thank you
Thank you so much for the clear explanation.
Very clear and to the point . . . .kindly make similar videos for each topic such as Decision tree classifier etc
Thank you so so so much for clearing the concepts.
Hi, your videos are awesome and easy to understand. Can you please upload the logistic regression, random forest, SVM and times series modeling videos with examples.
sure, will try to upload something soon
Thank you very much, and please upload with the dataset available on the internet, so that we can try on our own.
@@jaydeepraut820 did you try to search in Google for the dataset described, as she showed and explained it at the video timestamp @1:10 and following? As she describes this in the video, it is straightforward to search and find from Google.
thank you so much this video help me to understand the concept faster
Glad to hear that!
Do you have a guide on how to do multiple variables if they are non linear? Meaning we’d have to use a polynomial method with degrees?
This is sooooo Great!!!
GREAT WORK MEM👏👏👏👏👏
you are life saver
This is a very good video! Thank you very much!
Thank you so much! Super Clear!
Glad it helped!
Awesome video. keep on doing great
I really appreciate your work, however, I will like to point something I noticed out.
The scatter plot you created for your result at the end of the analysis where you had the y_test plotted against the y_pred seems inaccurate to me. plt.scatter(y_test, y_pred) is supposed to indicate that your y_test is on the x-axis while your y_pred. is to be plotted on the y-axis. I believe what you should do is showcase the y_test and y_pred on the y-axis while you use a common x-axis for the two on the same plot.
Two different ways of showing the same basic thing. y_pred vs. y_test will form a 45-degree straight line if the data is perfectly predicted, whereas y_test and y_pred both vs. x_test will show the points overlapping each other, if perfectly predicted. However, because this is multiple regression, there isn't a single x variable, so you would have several of the latter plots. These are useful for certain diagnostics. However, a very common first plot is y_test vs. y_pred. Plotting residuals vs. y_test or x_test (one at a time) are also common charts to make.
Thank you for your nice explanation, can you upload more videos other modelling, please? thanks you mam
thank you miss
Great explantion ............ur work must be admired
Nicely explained. Here, the difference is actually residual, right?
So useful!!! Thank You
you are the best!!
Amazing tutorial!
Really nice explanation. One question, in the final regression equation can we have 0 as coefficient for any independent variable or all variables will be assigned some non 0 values?
It should always have a non zero value because if it is zero it would mean that the specific independent variable is totally useless to predict the dependent variable, which never is the case.
Because you're estimating from a sample, it would be nearly impossible for a variable to have a coefficient of 0, even if there isn't any correlation between the dependent variable and the independent variable (within the context of your model) in the real process. If you ever get a zero coefficient, it's likely due to an error such as including a redundant feature, having perfect (multi)collinearity, including k dummy variables (and an intercept) for k cases, etc. This can be fixed by reducing the feature set (one at a time, of course.)
So good! this is much simpler explanation. I love it
Hi thank you for sharing. But I am wondering how do you get the actual linear regression equation with sklearn
very nice and easy teaching. Congrats.
Thanks!
Hi Megha, thanks for showing steps by steps. I have a question. Instead plotting the result by "Actual" and "Predicted", can we visualize the predicted vs. actual ''y" for each variable"x"? Can you please advise the codes? Thank you.
I liked the video, so I "Liked" and I "Subscribed"
Thanks, MN
Excellent💯👍
Very Nice Megha..Just if you give some explanation of functions which are using then it will be more clear. Nice Attempt!!!
Hello I have this problem how can I solve it?
(could not convert string to float: 'rainy')
Simply Brilliant !!
thanks a lot, please I want to know which algorithm is used (batch, stochastic...?) also can we show the cost function?
Mam plzz further model bi explain kardo...
Waiting for your further ML lecture mam
Should always perform a check for assumptions of Linear Regression when performing it. Otherwise, it can be dangerous or misleading.
Do you have any good material for guiding how to check those assumptions using the exercise presented in the video?
Keep it up and make videos on other models too mam.
I really like your well organized presentation structure!
Clear and Nice explanation....
Thank you
Great Ma'am
how to perform regression for the following equation: time= ( pressure)^a + (concentration of fuel)^b + (concentration of oxidizer)^c + exp(d/Temperature). Here, time is dependent variable on pressure, concentration of fuel, concentration of oxidizer and temperature. How to set non-linear model like the above equation
Great breakdown! Liked and Sub 👌
Excellent!
@ Megha, thank you for this. Have you done another video of a way to improve the model? If so, can you kindly share the link?
Thanks for the video. I built a model to predict rent across my country. The accuracy score is 43% or so. What can I do to improve it? I can send the script of needed. Thanks.
Could you please make a video on this topic from scratch without using the sklearn library to better understand the mathematics behind it or provide a useful link?
Please...
@Chandra Shakhar start with the website she shows in her discussion of sklearn: scikit-learn.org/stable/modules/generated/sklearn.model_selection.train_test_split.html , which leads you to the technical explanation at: scikit-learn.org/stable/modules/cross_validation.html#cross-validation
Very nice video. Thank you so much and Best of luck--Shakir, Bangladesh
Very well done mam,why didn't you do outlier treatment mam ?
Thank you so much. It was really helpful!!!
Really helpful tutorial! Thank you.
Hi Megha, thanks for tutorial. what if we have string in datasets (like types can be multiple strings not boolean e.g colors:blue, red, green ,black ) how we will convert it into float format cuz model only understands numbers.
brilliant video
Fantastic tutorial. Question I have is if I wanted to test my model on another dataset how could I do it once I have my coefficients and intercept? Best Wishes Peter
Hi Megha, thanks for this great video, very simple question may be for you, on your predicted values chart, is there a way to plot a straight line across those values?
Like simple linear regression?
Brilliantly explained. Can u make video for deployment of model to use with webpage/android or any programming language GUI through API . Also make such a beautifully explained video for ANN also.
hi ,
Thanks for this amazing video!
If we need to print the linear regression equation in the form of y= a+bx1+cx2 , how to do that?
Do `dir(ml)` and you will see a list of attributes. You'll find what you're looking for in ml.coefs_ and ml.intercept_.
Very nice explaination. Could you please tell how we can get equation for the predicted model?
Cleannn tutorial, the best of all, thx :))
very insightful many thanks for your impressive work
hello Megha, this is good one , easy for beginner ....kindly upload on clustering neural network also
Very good video... Also why did we import numpy library in the beginning.
Thanks, you can import it later too.
Mam I got an error ln[8]
Could not convert string to float:AT
How to fix it? Can anyone tell me?
Great explanation mam...just simple and smooth 😃..keep uploading videos
Very nice explanation. Thanks!! for the explanation
nice practical
Ma'am can we get regression for more than 2 independent variables w.r.t more than two dependent variable??
could you please tell more, using multiple regression which technique you follow? I mean OLS or else?
Mam, how do we predict multiple target values (y variables) with a single linear regression model?
This is such an amazin video. It helps me with my course. I subscribed. Could you please make a video on Ploty Dashboard?
Nice explanation 👍🏻
Thank you!
simple explanation... Thank you mam
Thanks @Dhanesh
If we have two outputs in multiple linear regression model. What are the steps ?
After define x & y print statement should be came in " str " how it is possible if all dataset and format as it is copy .🤔
Superb explanation madam thank you
@Megha thank you so much for clear explanation. Much appreciated. can you pleased help to create video on how to improve model further.
sure, I will try to upload a new video soon
thx for video a lot. But what about the prediction of finding most minimum value in according to independents variables. How can we find it? thanks in advance
Very useful
THANKYOU GREAT VIDEO
Thank you so much. This is very helpful.
Thanks a lot..this video has helped me a lot in my project❤❤❤❤
Glad I could help :)