Build a ML Classification Model in 12 Lines with PyCaret

[@user-or7ji5hv8y]

I think one on anomaly detection or regression with PyCaret would be interesting.

[@sebastiansanchezmendoza7523]

Please a full course would be amazing!!

[@NicholasRenotte]

Got it all coded up! Just waiting for a cool down in work to start recording.

[@ameerazam3269]

Just before I'm searching for PyCaret Now i got amazing Sir

[@NicholasRenotte]

Ayyeeee! Yes!

[@rubbercable]

Great Work. Small byte-sized intros for new stuff certainly is a good formula. It works well.

[@NicholasRenotte]

Thanks, figured I'd push out some shorter vids as well, glad you liked them. Was getting exhausted from the long ones!

[@gaaya32007]

Could you please make a video on Feature Selection through Pycaret

[@michiel4195]

Hi is there anyone how can tell me how I can load the output array into a csv file that combines it with the correct rows? And save it as a new csv

[@yourassignment6046]

Wow, I didn't go through the video yet , but I like your 'fair' impact to the students who are suffering from Data science projects, Thank you men!...

[@NicholasRenotte]

Anytime man! Let me know how you go!

[@saikiran7652]

Once try "detecto" which can be used for custom object detection within 10 lines of code.

[@NicholasRenotte]

Oooooh, will check it out!

[@JBAhmad-vu2lr]

The categorical_features parameter gives me an error: cat_features =['sex',' cp', 'fbs', 'restecg', 'exang', 'thal'] experiment = setup(df, target = 'condition', categorical_features = cat_features) The error is: ValueError: Invalid value for the target parameter. Column target not found in the data. When I change those two lines to: cat_features =['sex',' cp', 'fbs', 'restecg', 'exang', 'thal'] experiment = setup(df, target = 'condition') ... no errors Could you please explain why?

[@JBAhmad-vu2lr]

Nick, could you please do a video on "shap" package?

[@ziadahmed7439]

How to predict and save top best 3 models ??

[@factsthatfizz]

Does it replace the traditional way to modelong or it could be just to select best models?

[@144digital]

Thank you so much for this... Please I have a question... In neuralprophet prediction, the RegLoss column that follows the MAE column after the training; what is it? Is it a measure of loss? If so, how does it differ from the SmoothL1Loss and MAE?

[@NicholasRenotte]

Yup, spot on. RegLoss is regularised loss (loss with an applied penalty), SmoothL1Loss is another measure of loss with an L1 penalty applied. MAE is mean absolute error!

[@Progg95]

Great tutorials! It would be great if you could show the finished project and its outputs at the beginning of the video so we know what the result looks like :)

[@NicholasRenotte]

Anything in particular? Like the input data and the output data?

[@Progg95]

​@@NicholasRenotte No, just a short demo of what the code does. I didn't mean this video in particular, but in general. For example if you do some ML with the webcame that detects objects in you flat you could show a short demo of what this looks like at the beginning of the video. This motivates the video, because the viewers see what the end result will look like.

[@ravikumarpande4143]

make a complete series on pycaret thank you!

[@NicholasRenotte]

You got it!

[@savirarahmania6587]

This is amazing video. I wanna ask, after we compare model and get the best model to use. How we fix it or i mean how we lock that model, because when i running to compare again for the 2nd Times or 3rd times, the results of compare model is different and the results of best model is also different. So my question is how to lock the best model ?

[@syaduinotech3681]

Thanks sir. I got problem in pycaret installation. Can you share the video about pycaret and how to install it and use it. I am using spyder. Sorry if the question is newbies standard

[@alexandertsikhun7733]

hey, great video! I have question - why in the list of all models auc for ridge clf = 0, but when you write "predict_model(best_model)" auc=0.82??

[@muhammadghazy9941]

I am waiting more of pycaret content from you and i am subscribing

[@NicholasRenotte]

You got it!

[@chiranshuadik]

Hi Nicholas! Can you suggest a method to implement 'quick reply recommendations' for chat applications? (We can see those in linkedin comments and messages). I need to do that in python or javascript. Thanks!

[@NicholasRenotte]

Like autosuggest responses? Hmm, would need to dig into it!

[@chiranshuadik]

@@NicholasRenotte yes (like Gmail, linkedin does) , I was thinking to use rasa intent based chatbot to generate the responses. Do you think that would be a good way to go? Thanks :)

[@NicholasRenotte]

@@chiranshuadik ah got it, haven't used Rasa much before but I've heard good things!

[@chiranshuadik]

@@NicholasRenotte thanks I'll try that! Will let you know how that goes

[@ingeday]

@@chiranshuadik Did you do anything?

[@saivarma3406]

May I know which alogrthim u used to make sign language detection

[@NicholasRenotte]

From which video?

[@linkinmaniac99]

Type error : data passed must be of type pandas.DataFrame Yes my dataset is of .dat format, what should I do?

[@NicholasRenotte]

Check this out: stackoverflow.com/questions/37956344/reading-and-doing-calculation-from-dat-file-in-python

[@PUBUDUCG]

This is great and they do have anomaly detection too. . . . however its strange they do not have object detection?. You have very broad knowledge in this field, how is this compared to Tensorflow? (Not the code complexity)

[@NicholasRenotte]

Heya, from what I can see it's focused on tabular data. TF is more of a free for all deep learning library.

[@prashanthkatakam6336]

Great video as always would love to see more of pycaret waiting for the full tutorial

[@ronakttawde]

can please add the code on how to select the other models than best model drawn in the code? thank you for this video.

[@NicholasRenotte]

Heya @Ronak, can use the create_model function: pycaret.org/create-model/

[@arianklows4493]

Hi Nick, Thank you for uploading this video. I tried it and ended up having this error. >>> experiment = setup(df, target='target', categorical_features=cat_features) IntProgress(value=0, description='Processing: ', max=3) Initiated . . . . . . . . . . . . . . . . . . 09:16:15 Status . . . . . . . . . . . . . . . . . . Loading Dependencies Initiated . . . . . . . . . . . . . . . . . . 09:16:15 Status . . . . . . . . . . . . . . . . . . Preparing Data for Modeling Initiated . . . . . . . . . . . . . . . . . . 09:16:15 Status . . . . . . . . . . . . . . . . . . Preparing Data for Modeling Traceback (most recent call last): File "", line 1, in File "/usr/local/miniconda3/lib/python3.9/site-packages/pycaret/classification.py", line 580, in setup return pycaret.internal.tabular.setup( File "/usr/local/miniconda3/lib/python3.9/site-packages/pycaret/internal/tabular.py", line 1214, in setup fold_generator = StratifiedKFold( File "/usr/local/miniconda3/lib/python3.9/site-packages/sklearn/model_selection/_split.py", line 644, in __init__ super().__init__(n_splits=n_splits, shuffle=shuffle, random_state=random_state) File "/usr/local/miniconda3/lib/python3.9/site-packages/sklearn/model_selection/_split.py", line 296, in __init__ raise ValueError( ValueError: Setting a random_state has no effect since shuffle is False. You should leave random_state to its default (None), or set shuffle=True. >>> Tried setting shuffle-=True, but no luck... >>> experiment = setup(df, target='target', categorical_features=cat_features, shuffle=True) Traceback (most recent call last): File "", line 1, in TypeError: setup() got an unexpected keyword argument 'shuffle' Could you please help me? I've no idea, what this means?

[@priyabratapanda1216]

Thank you @Nicholas. Looking forward to full course on Pycharet 🙏🙏

[@NicholasRenotte]

YESSS! Awesome!

[@AiykRichie]

I really like the course, am getting started with automl and mlops, and found your course really helpful. though there is one challenge i faced why trying to follow along. I got this error --------------------------------------------------------------------------- ModuleNotFoundError Traceback (most recent call last) Cell In [2], line 1 ----> 1 import pycaret.classification File ~/miniforge3/envs/bsfpd/lib/python3.8/site-packages/pycaret/classification.py:10 7 import pandas as pd 8 import numpy as np ---> 10 import pycaret.internal.tabular 11 from pycaret.internal.Display import Display, is_in_colab, enable_colab 12 from typing import List, Tuple, Any, Union, Optional, Dict File ~/miniforge3/envs/bsfpd/lib/python3.8/site-packages/pycaret/internal/tabular.py:33 31 import pycaret.internal.patches.yellowbrick 32 from pycaret.internal.logging import get_logger ---> 33 from pycaret.internal.plots.yellowbrick import show_yellowbrick_plot 34 from pycaret.internal.plots.helper import MatplotlibDefaultDPI 35 from pycaret.internal.Display import Display, is_in_colab File ~/miniforge3/envs/bsfpd/lib/python3.8/site-packages/pycaret/internal/plots/yellowbrick.py:8 6 from pycaret.internal.logging import get_logger 7 from pycaret.internal.Display import Display ----> 8 import scikitplot as skplt 9 import matplotlib.pyplot as plt 12 def show_yellowbrick_in_streamlit( 13 visualizer, outpath=None, clear_figure=False, **kwargs 14 ): ModuleNotFoundError: No module named 'scikitplot'

[@AiykRichie]

need to mention I am running mine in a miniforge conda environment, host pc is a mac with m1. would appreciate any assistance

[@peralser]

Thanks for sharing this content. Appreciate your time and support.

[@ramratan2004]

Really great content. Its boosted my confidence on ml model building. Plz keep this up..👍

[@expenditorkaundachungu9435]

great tutorial!! Well demonstrated.

[@hemachandiran3984]

Nick,can we use pyCaret for Regression?

[@NicholasRenotte]

Sure can! There's a regression module!

[@espua72]

what a nice introductory for pycaret, so helpful, and so easy to follow. Wish you could make few more video on this topic

[@NicholasRenotte]

Definitely more coming!

[@jaehwilee9614]

It seems that the data is not available on Kaggle anymore. :( But nevertheless, great video. Thanks!

[@anoapenshapiro2733]

Hi Nick, this is very interesting. You explained pycaret very nice and simple. Can you please do a video on deploying this to both local machine and cloud? I heard pycaret could deploy with few lines of code as well

[@henkhbit5748]

as always a clear introduction. btw: u installed shap but did not used it for explaining the feature importance 😎 Interested alos in anomaly detection with autoencoder and credit fraud data and shap...

[@NicholasRenotte]

Thanks a mil @Henk, it was a dependency for PyCaret but yeah I didn't go too deep. Pumped for the race this w/e? Also I made a huge course on PyCaret last year but never released it, might need to get back to it.

[@henkhbit5748]

@@NicholasRenotte Would be nice if u can give more elaborate course on Pycaret. I think Pycaret is good ML library with a low learning curve where many tedious tasks like encoding categorical data as u showed will be transformed automatically.

[@jaaaaaang]

thanks for a good video !! I have a question. when I am using pyCaret, It takes more time than using deep learning like tf, keras. I think it might be using GPU. are there some tips or options for saving time when we use pyCaret? have a good day!!

[@okanpasa]

Very useful thank you very much!! Just a question : When it compares the models, does it make it by comparing the results in test data or train data? (You didn't split the data)

[@NicholasRenotte]

Heya @Mert, good question. The PyCaret library does a native ttsplit, the default train partition size is 70%. More here: pycaret.org/train-test-split/

[@okanpasa]

@@NicholasRenotte Thank you very much, i appreciate very much your approach!

[@soranlanguage8302]

Plz make a pycaret series

[@NicholasRenotte]

You got it!

[@lebedev_michael]

Awesome!!! Respect from Russia)

[@muhammadghazy9941]

this is surprisingly easy

[@NicholasRenotte]

I know right? Compared to my normal workflow, it's wayyyyy more straightforward!

[@jaygadhavi6786]

Thank you so much for this video, and can you please make a video on best project for top machine learning algorithms/models?

[@m.niharikareddy9249]

Ur explanation was really good & knowledgeable ʘ⁠‿⁠ʘ

[@swapnil6792]

Thank you so much. Please do make videos on Pycaret... It was really helpful

[@NicholasRenotte]

You got it @Swapnil!

[@AkshatKalra-gw5cx]

Thank you for this super comprehensibe explanation. Can you also make another video for using pycaret to train a models for let's say Customer Loan Churn dataset, which has a majority of numeric variables.

[@user-or7ji5hv8y]

what's unclear is why you didn't pass 'experiment' to compare_models() as an argument? how does compare_models() know what to do its training on?

[@NicholasRenotte]

It's running on the existing experiment @C :)

[@xyliusdominicibayan6215]

Hey man can you do a tutorial for Facial Recognition using React Native Mobile Application. Thank you!

[@NicholasRenotte]

Working on the beginnings of this atm!

[@DeonBands]

Love your work, learned a lot. I am sure the code worked when you demonstrated it. Up to this point all your content works first time. This one left me stranded, indicating an error. "Setting a random_state has no effect since ........" I went on various forums and this is an issue that only popped in recently. It is also an issue that is open on the github repo. If I was just starting out on python I would have been bombed and would have thought the issue is with me and my skills. Just a heads-up it is not with skills or your content. It is an issue that recently popped in and last commented on in August 2021. Hope it get's solved this tool has huge potential, which is why I took the time to comment. I never comment on content. Again love your work.

[@user-or7ji5hv8y]

Perfect timing.

[@lasnamiwalid3633]

Well explained I really appreciate it Thanks

[@rajaardin2894]

I like all of your tutorial about python. Can you make one tutorial for beginner about making games or something ? 😇😇

[@NicholasRenotte]

Got something coming for beginners soon!

[@rajaardin2894]

@@NicholasRenotte ahhh thanks. I'll be waiting for that 😇😇😇

[@AfghanGoldenOldies2]

Can you please do your next project in Car Belt Detection. Many Thanks from Afghanistan :)

[@NicholasRenotte]

Will look into it!

[@dramekandya4918]

Very Smart...Wonderfull Thanks a lot!!!

[@NicholasRenotte]

Thanks so much!

[@clovenbeast5183]

Why don't you have more subscribers?

[@NicholasRenotte]

KZbin doesn't love me 😂, nah I kid, I'm just grateful I got some awesome subscribers watching these :) just spreading some tech love!

[@riffaxelerator7299]

Its only a matter of time!

[@nilszimmermann7102]

where SHAP? For real though, many thanks

[@zhaowang8913]

Awesome!

[@NicholasRenotte]

Thanks a bunch @Zhao Wang!

[@utsavanand566]

💯💯

[@NicholasRenotte]

❤

[@winniepooh2533]

Nice!

[@NicholasRenotte]

Thanks a bunch @Muntasir!

[@bhargavraghuram]

❤❤

[@jagatdhiman9490]

Jiko

[@NicholasRenotte]

🤗

[@King_Fury]

a less heap of video with whole heap of information and whole heap of coding

[@jagatdhiman9490]

Ii

[@NicholasRenotte]

🤗