Live-Feature Engineering-All Techniques To Handle Categorical Features - Day 4
Dataset link: drive.google.com/file/d/1hJgt...
Live Streaming Playlist: • Live stream playlist
Telegram link: t.me/joinchat/N77M7xRvYUd403D...
github link: github.com/krishnaik06/Featur...
Join the Ineuron Affordable course
ineuron1.viewpage.co/Deep-lea...
Please donate if you want to support the channel
Gpay: krishnaik06@okicici
Please join as a member in my channel to get additional benefits like materials in Data Science, live streaming for Members and many more
/ @krishnaik06
Please do subscribe my other channel too
/ @krishnaikhindi
Connect with me here:
Twitter: / krishnaik06
Facebook: / krishnaik06
instagram: / krishnaik06
Пікірлер: 65
i AM TRUELY TELLING THIS SIR, HOW MUCH GOOD WORK YOU ARE DOING FOR THE PEOPLE WHO ARE IN NEED LIKE ME, EVEN PEOPLE CANNOT IMAGINE IT. THANK YOU SO SO SO MUCH FOR YOUR EFFORT, WHERE WE ALL WERE WHERE YOU BROUGHT US THROUGH YOUR CHANNEL, CANT SAY IN WORDS THANKS AGAIN SIR.
Great sir!You have been doing a great work to give us good knowledge for many days..Thank you for giving us such great knowledge in all areas of AI.
Thankyou Krish Sir for conducting these sessions, they are very helpful .
He is a living legend.
You picked up Feature Engineering on the 7th heaven 😎😎A Big THANK YOU...🙇🙇🙇
Thank you for your efforts your time sharing your knowledge, thank you
Great work krish...keep rocking like this
thank you sir !!!! for making available this "one stop platforms" . and doing time investment on rusted train like me...
Thanks for all the live videos sir
change "data['weekday']=data['Day'].dt.weekday_name" to "data['weekday']=data['Day'].dt.day_name()" if facing *ERROR*
@thepresistence5935
2 жыл бұрын
all are facing error
@sanketjaisingpure7028
2 жыл бұрын
Ya 👍
hats off, amazing content
really Video is very helpful
Thank you krish🙏
Thank you sir 🌟
Thank u, sir, can u post the video for the feature selection method for unsupervised learning of categorical features
more advanced feature engineering videos sir...great explanation!!!!!
instead of hardcoding the dictionary we can use this:- import calendar dictionary=dict(zip(calendar.day_name,range(1,8)))
very useful session
Excellent
I found this code snippet for quick dict comprehension for days of week. import calendar >>> d=dict(enumerate(calendar.day_name)) >>> d
we can use label encoding with sklearn library
Hi Krish, i have question. We have one more technique to impute high cardinality variable i.e. picking top 10 most frequent categories. So in such case can we either of them count or frequency encoding or one hot encoding of top categories ?
Sir you have explained the concepts in very effective way , but i have one question, If you we do ordinal encoding with respect of target variable. What about the test data .. We will not have target variable. How to deal with this scenario.
finished watching
how to know which encoding is best suited where when we deal with data set for novice coder it is difficult too identify what to apply and which one is best
Good Morning Krish, I have a request, Please try to put the links used in the Video in the "Video Description", this will help while watching the videos after the Live stream. as we cannot access the Stream chats after the stream is over.
Continue sir feature engineering
I new to Machine learning I love the way to teach may I know from which playlist I have to start . This helps alot people to learn some basic to understand better live streaming
@akashubale5199
3 жыл бұрын
Bro you can start with his Machine Learning Playlist.
val = list(data['weekday'].unique()) dicts = {} keys = val[::-1] ran = len(keys) i = 0 while i dicts[keys[i]] = i+1 i= i+1 #to Complete above assignment using loop
I'm confused, please guide me. can we use frequency encoding in both type of features (i.e. Nominal or Ordinal)? if yes then how it will affect to nominal type of feature. here you use frequency encoding for nominal type of feature(i.e. country name), I think after that this nominal feature is reacts like ordinal feature because, labels with higher frequency get higher weightage and lower frequency get lower weightage.
Sir plz check this issue that when I am handling imbalanced data it shows the attribute error of validating data so can u plz make a video on this
HI sir loved your videos and I am your student for a long time sir I have one doubt when we do target guided ordinal encoding on the train dataset it is fine but sir if we have a test dataset which is untouched and I want to do the same encoding because While testing my model it is giving me shape mismatch error so can you help me with this..?
Friends, What if we have a column (say "Location") which consists of more than 1000 categorical variables?? FYI, this column is an independent variable and one of the important parameter for predicting the label. Answer pls.... Thanks in advance
after enoding we have to drop the previous feature or not ?? please help
Yes
if we have regression problem our output will be continuos that time how to do the encodings(target guided,mean)?
Count encoding: If the frequencies are not close together, it will produce numerical column which you can't use. Like 20000, 14000, 3000, 2000, 250. So again standardize ?
I am working on a project named Face mask detection supervised data.There has 2folders namely With mask and Without mask , Q)How to read all images from both folders and classify
Why weekday is considered an ordinal category?
Guys, Please do a video on the Story of a person without a degree become Data Scientist. I don't care whether it is a Successful one or not.
sir make videos of feature selection
🙏
Hi krish, need to understand one thing. excel have 3 columns and if we perform one-hot encoding, X1 create 10 new columns, X2 create 9 new columns and X3 create 6 new columns. why? I thought all 3 will create 10 each new columns. Please help me to understand anyone.
Yeah
dt.weekday_name is now deprecated and instead of that we can use dt.day_name()
@kumariginka4400
3 жыл бұрын
Thank you so much
@kumariginka4400
3 жыл бұрын
Amit i reallu searched for weekday _name thanks alot
@kumariginka4400
3 жыл бұрын
I also use day _name but putting () may gives results lastly
@amitKumar_11
3 жыл бұрын
@@kumariginka4400 😇 your welcome
how the weekday can be ordinal features... we can't campare weekdays to each other .. yes of course there is sequence but i think doing ordinal encoding on weekdays isn't good option , in this case our model will give more weight to the Sunday 🌞 as it's mapped with 7️⃣
In colab this command is not working which is data['Day'].dt.weekday_name It is giving me error as AttributeError Traceback (most recent call last) in () ----> 1 data['Day'].dt.weekday_name AttributeError: 'DatetimeProperties' object has no attribute 'weekday_name'
where can i find this train_set datasety...anyone please
@rushabhwalke4205
3 жыл бұрын
it will give you error if you load it on your browser but it will work fine when you will load it in your Jupyter notebook with read_csv() function Happy DS!
whoever got this error, "DatetimeProperties' object has no attribute 'weekday_name," use data['weekday'] = data['Day'].dt.strftime("%A") it will work
@sudharsanb9391
3 жыл бұрын
thank you
@MalluQA
3 жыл бұрын
@@sudharsanb9391 Thanks Buddy
@mayanktripathi4u
3 жыл бұрын
week_day may not work for Python 3 + versions.. if you are using Python 3 + please use day_name Code : data['Day'].dt.day_name()
feature extraction videos
starts at 00:43 for those coming for the juice
he syntax you were trying to use data['Date'].dt.weekday_name is for PANDAS