Naive Bayes Classifier in Python (from scratch!)

Let's create a Naive Bayes classifier with barebone NumPy and Pandas! You'll learn how to deal with continuous features and other implementation details.
#machinelearning #datascience #python
For more videos please subscribe -
bit.ly/normalizedNERD
Join our discord -
/ discord
Bayesian classifier concept video -
• The Math Behind Bayesi...
Source code -
github.com/Suji04/ML_from_Scr...
Like my work? Support me -
www.buymeacoffee.com/normaliz...
Data source -
www.kaggle.com/merishnasuwal/...
Facebook -
/ nerdywits
Instagram -
/ normalizednerd
Twitter -
/ normalized_nerd

[@HaiderAli-hp6tl]

the number of subscribers to your channel does not do justice to your content. This is such quality educational content. Keep it up, man.

[@sayantansadhu6380]

The from scratch series in this channel is the best !!

[@NormalizedNerd]

Thanks bro!

[@secondarypemail7181]

Thanks man,your effort to make algorithms from scratch is just on another level.Your effort is much appreciated👍

[@user-nd6ll6en9j]

ممكن تبعثلي لكود

[@caljohn1475]

You're a legend my dude, thanks so much for explaining this

[@glennbroyne7316]

Fantastic video, very well explained!

[@tarabalam9962]

this gave a lot of clarity , thanks

[@cheatforlife]

Thank you for opening up new horizons for me

[@amitkumarghosh]

I can't believe this man is doing this for free, thank you brother

[@lucascojot7226]

Super high quality videos! I'm surprised you have 8K and not 800K... Keep it up!

[@NormalizedNerd]

Thanks a lot! Keep supporting :D :D

[@pietrogarofalo2957]

Man, u save my life ty very much.
Use sklearn is too easy, justify why u decide to use Naive and why u can use it is the very important thing, keep it up man .
( excuse me for my bad english )

[@kushagrakumar9576]

Excellent video. Keep up the good work 🙂

[@NormalizedNerd]

Thank you! Will do!

[@averyiorio4337]

amazing content and fantastic explanations

[@NormalizedNerd]

:D

[@jorgetaboada6500]

Sorry, again I do understand now, and also I apply in my work with excellent results, Thanks!

[@NormalizedNerd]

Great to hear that!

[@sukhmanpreetsinghsandhub2042]

Awesome video.

[@caljohn1475]

@normalized Nerd How do you make a prediction with this using specific values?

[@AbhishekS-cv3cr]

You can also ignore mean_radius feature, since it has some correlation with other features

[@vishwajeetdeulkar3862]

Hi, I am getting error as "index 29 is out of bounds for axis 0 with size 29" for this statement likelihood[j] *= cal_gaussianLikelihood(df,features[i],x[i],Y,labels[j]), any solution?

[@minasahebi1474]

perfect!

[@tudorjeverdan]

I was importing a mysql dataframe, I was importing string elements and it resolved them into objects,
data = pd.read_sql_table("ai_learning", engine)
columns_to_convert = ["Products", "feedback", "blog", "diagnosis"]
data[columns_to_convert] = data[columns_to_convert].apply(pd.to_numeric, errors='coerce')
data = data[["Products", "feedback", "blog", "diagnosis"]]
This is how I fixed it if anybody was getting the same outputs.

[@muhammadirsyaduddin6597]

Hello! Is it possible to add the multinomial in the code? Thank you.

[@yashicasaun]

When you are checking for gaussian curve, shouldn't you have filtered for different diagnosis and then check if the curve fits?
Because now, we see the data fits gaussian. But we then change the data and only take a subset and then fitting the curve
Thanks for the great video.

[@dimamedvedev4124]

Great lesson. But still don't understand how we get array with posterior given certain set of features befor using np.argmax?

[@hajraali1205]

I did not understand the output, we were detecting the cancer patient, but in out put there are two matrix and accuracy data so which is which.

[@placement10lpa36]

can someone explain me the guassian distribution part

[@LucasPiorino]

What 36 and 74 means at 13:49?

[@Mustistics]

Damn, I was hoping for a SKlearn tutorial!

[@NormalizedNerd]

opps

[@teetanrobotics5363]

SVMs,Random Forest and gradient boosting left in the playlist

[@NormalizedNerd]

Humm...I'll add them eventually.

[@UdayKumar-yv7ej]

Bro for this code how to convert string to float bro

[@shortyTalks]

😃Bro thx for the nice explanation. Are you using a theme for vs code, cuz all the colours in your systems are looking damn good

[@abhisheksinghyadav4970]

atom material theme

[@pritammukherjee3972]

Nice video

[@NormalizedNerd]

Thanks!

[@alexchristian220]

likelihood = [1] * len(labels), post_prob = [1] * len(labels)
what this above code actually do?
And also how can I work this code on tennis.csv dataset?

[@alexchristian220]

bhai iska explanation bta do.

[@NormalizedNerd]

That's just a python shortcut.
arr = [1] * n
this will assign a list of n 1's to the variable 'arr'. I used this to initialize the likelihood and posterior probability lists.
The tennis dataset contains only categorical variables so just relabel them to 0,1,2, etc. and apply approach 2 given the video.

[@alexchristian220]

@@NormalizedNerd bhai can you tell me what is the name of the python shortcut? and how likelihood = [1] * len(labels), replaced with likelihood[b].?

[@jorgetaboada6500]

Sorry, but I do not understand who is "df" when you def a function because you have never defined. I will appreciate your explanation

[@Joemama-jh5go]

DataFrame, just means the data