Thank you, been struggling for a couple of days (after even using the sklearn library example) to understand ROC curve for data mining / machine learning class.
Sir with help of this i can predict knn auc score but after that when i write r_fpr, r_tpr, _ = roc_curve(y_test, r_probs) the above code shows error "y_true takes value in {'Abnormal', 'Normal'} and pos_label is not specified: either make y_true take value in {0, 1} or {-1, 1} or pass pos_label explicitly" how do i solve it
Hi Ishmam, I've modified the Jupyter notebook described in the video tutorial to also include knn and it is available on the Data Professor GitHub at github.com/dataprofessor/code/blob/master/python/ROC_curve_kNN.ipynb
@@DataProfessor Sir i actually i did exactly like that and after that i got the error.I was actually reading data with pandas from a csv file.i also used label encoder to ensure there are no strings and still after that got the error.
How to do that for molecular docking ligand enrichment study.. I have 2 csv files. One contains docking score of true inhibitors and another decoy inhibitors. Plz guide
Thanks for this informative video. Sir, I have a question. If I want to apply this code to display roc curve of ML models for five different class of sentiments, what should be included in this code? please reply as quickly as possible.
Thanks for watching, for different class labels, you can check out this code example and adapt for your project scikit-learn.org/stable/auto_examples/model_selection/plot_roc.html#plot-roc-curves-for-the-multilabel-problem
thank you for the video. I got this error while plotting: ValueError: multi_class must be in ('ovo', 'ovr'). What can I do. And is roc is only valid for data with 2 class labels?
Hey , Very informitive and helping video. I have one doubt though, does the Y attribute for KNN classification need to be binary in order to plot in roc curve?
@@DataProfessor sir is svm doesn't support this function predict_proba and can I use dataset which I can read in python that is CSV file instead of synthetic dataset to generate roc curve
@@DataProfessor sir is svm doesn't support this function predict_proba and can I use dataset which I can read in python that is CSV file instead of synthetic dataset to generate roc curve
@@HimaniChauhan To use CSV file instead of synthetic dataset, you can read it in using: import pandas as pd df = pd.read_csv("data.csv") X = df.iloc[:,:-1] Y = df.iloc[:,-1] You can then use X and Y as your input argument for model building
Hi, please refer to this for coding example scikit-learn.org/stable/auto_examples/model_selection/plot_roc.html?highlight=roc#plot-roc-curves-for-the-multilabel-problem
Warawat, thanks for your comment. For the moment, I recommend the following 2 books. For the principles, I would recommend "The hundred-page machine learning book" themlbook.com/ For practice, I recommend the "Python Data Science Handbook" jakevdp.github.io/PythonDataScienceHandbook/
Sir I am new to this field. I want to know the meaning of these codes. r_probs = [0 for _ in range(len(Y_test))] rf_probs = rf.predict_proba(X_test) nb_probs = nb.predict_proba(X_test) Is there any other way of writing this?
Hi, I believe the answer you're looking for lies in this Stack Overflow post stackoverflow.com/questions/19984957/scikit-learn-predict-default-threshold?rq=1
@@shwetaredkar734 Sure, I've previously made a video and drawn an infographic on the overview of machine learning model building. Explanation on the external set (test set) is also provided. th-cam.com/video/BOk1hlCPW0c/w-d-xo.html
@@shwetaredkar734 Thanks for your comment. To apply the trained model to make prediction on the class label using clf.predict(X_test). Please see one of the recent video th-cam.com/video/XmSlFPDjKdc/w-d-xo.html
Yes you can, check out the pythonmeta library (pip install PythonMeta) which allows you to make forest plot You can also try making one using their web demo pymeta.com/
Really clean code. I read it and understood immediately all of it. I was only bothered by the X,Y instead of the more conventional X,y.
Thanks for watching and for the comments 😃
Thank you, been struggling for a couple of days (after even using the sklearn library example) to understand ROC curve for data mining / machine learning class.
Awesome, hope it worked out.
Watching this trying to get through my internship because they expect us to know all this stuff I guess lol. Thanks brah.
Thanks Robin for watching, good luck in the internship 👍
Your code is so well organized and clean. Thank you very much!
Awesome content. I wish you get a million subscribers within 1 year
Wow, thanks for the kind wishes 😃
Sir with help of this i can predict knn auc score but after that when i write
r_fpr, r_tpr, _ = roc_curve(y_test, r_probs)
the above code shows error "y_true takes value in {'Abnormal', 'Normal'} and pos_label is not specified: either make y_true take value in {0, 1} or {-1, 1} or pass pos_label explicitly"
how do i solve it
Hi Ishmam, I've modified the Jupyter notebook described in the video tutorial to also include knn and it is available on the Data Professor GitHub at github.com/dataprofessor/code/blob/master/python/ROC_curve_kNN.ipynb
@@DataProfessor Sir i actually i did exactly like that and after that i got the error.I was actually reading data with pandas from a csv file.i also used label encoder to ensure there are no strings and still after that got the error.
How to do that for molecular docking ligand enrichment study.. I have 2 csv files. One contains docking score of true inhibitors and another decoy inhibitors. Plz guide
Thank you so much😊😊😊😊
THANK YOU SO MUCH FOR THE HELP!
Thank you for encouraging , tutorial is very good
Thanks thanks thanks. I love this chanel
Glad you enjoyed the contents here! Would greatly appreciate if you could spread the word, Thanks!
Thank you very much sir. Bless you
You are most welcome
Thank you for explaining and showing it!
A pleasure. Glad it was helpful!
Great job! Thank you for making it simple.
A pleasure! Thanks for watching!
IndexError: too many indices for array: array is 1-dimensional, but 2 were indexed --- I got this error while using for my code. plz tell the solution
may god bless you. This video really saved me ^ _ ^
I'm so glad! Thanks for watching!
Thank you! Very incredibly helpful, sir :)
Thanks for this informative video. Sir, I have a question. If I want to apply this code to display roc curve of ML models for five different class of sentiments, what should be included in this code? please reply as quickly as possible.
Thanks for watching, for different class labels, you can check out this code example and adapt for your project scikit-learn.org/stable/auto_examples/model_selection/plot_roc.html#plot-roc-curves-for-the-multilabel-problem
How to create this type of chart in .CSV file in matplot library ? Is possible means share the link
Amazing! Thank You!
Walton, Thanks for the kind words!
Guys I found the threshold , its 0.8 for random forest now how do I make the necessary changes in my model?
thank you for the video.
I got this error while plotting:
ValueError: multi_class must be in ('ovo', 'ovr').
What can I do. And is roc is only valid for data with 2 class labels?
thankyou sooo much sir
Hey , Very informitive and helping video. I have one doubt though, does the Y attribute for KNN classification need to be binary in order to plot in roc curve?
Yes, the Y needs to be qualitative/categorical labels for ROC to work.
what to do if predict_proba is not available when probability=false
Can you try setting probability=True
@@DataProfessor sir is svm doesn't support this function predict_proba and can I use dataset which I can read in python that is CSV file instead of synthetic dataset to generate roc curve
@@DataProfessor sir is svm doesn't support this function predict_proba and can I use dataset which I can read in python that is CSV file instead of synthetic dataset to generate roc curve
@@HimaniChauhan There is a workaround that may work. Can you have a look at this link: www.kaggle.com/c/home-credit-default-risk/discussion/63499
@@HimaniChauhan To use CSV file instead of synthetic dataset, you can read it in using:
import pandas as pd
df = pd.read_csv("data.csv")
X = df.iloc[:,:-1]
Y = df.iloc[:,-1]
You can then use X and Y as your input argument for model building
can i plot roc curve from saved model ?
Thank you Sir
There's only 6 thresholds from roc_curve function, how can I add more?
I got an error
ValueError: could not convert string to float: 'data'
How can you help?
I want to ask you a favor. can you explain how to plot ROC curve using Image Generator(flow from directory) in classification multi classes?
Hi, please refer to this for coding example scikit-learn.org/stable/auto_examples/model_selection/plot_roc.html?highlight=roc#plot-roc-curves-for-the-multilabel-problem
Thank you so much!
Thanks you for video. Could your please suggest a book or video about machine leaning to learn?
Warawat, thanks for your comment. For the moment, I recommend the following 2 books. For the principles, I would recommend "The hundred-page machine learning book" themlbook.com/ For practice, I recommend the "Python Data Science Handbook" jakevdp.github.io/PythonDataScienceHandbook/
@@DataProfessor Thank you for suggesting the books.
can i plot ROC with confusion matrix only?
I would plot the ROC and show the confusion matrix as a table.
No, you'll need the probabilities your model assigns for the classes and the actual class to plot the roc curve.
Thank You
You're welcome
Sir I am new to this field.
I want to know the meaning of these codes.
r_probs = [0 for _ in range(len(Y_test))]
rf_probs = rf.predict_proba(X_test)
nb_probs = nb.predict_proba(X_test)
Is there any other way of writing this?
I want to know how to apply it in real dataset available out there in the internet. Thank you and sorry for my lack of knowledge.
Nice video
Thank you 😊
Good Evening sir, I just want to know how to adjust cut off / threshold of a logistic regression model. Thank you, So much sir.
Hi, I believe the answer you're looking for lies in this Stack Overflow post stackoverflow.com/questions/19984957/scikit-learn-predict-default-threshold?rq=1
Could you please show the same example using 10 fold-cv instead of train test split. Thanks
Thanks Shweta for the suggestion. I will make a video shortly.
I’m working on making a Jupyter notebook of it, I’ll paste the link here when I’m finished
@@DataProfessor Okay. Thank you. Also, if possible please educate us on how to test an independent/external dataset on the trained model.
@@shwetaredkar734 Sure, I've previously made a video and drawn an infographic on the overview of machine learning model building. Explanation on the external set (test set) is also provided. th-cam.com/video/BOk1hlCPW0c/w-d-xo.html
@@shwetaredkar734 Thanks for your comment. To apply the trained model to make prediction on the class label using clf.predict(X_test). Please see one of the recent video th-cam.com/video/XmSlFPDjKdc/w-d-xo.html
how to plot false negative rate with false positive rate?
How to plot a ROC curve using qPCR data?
Can we make a forest plot???????????
Yes you can, check out the pythonmeta library (pip install PythonMeta) which allows you to make forest plot
You can also try making one using their web demo pymeta.com/
I LOVE YOU
Are you dutch?