test accuracy for : simple linear regression : 0.76 svm regression: -0.11 random forest regression: 0.803 so our choice right now can be to plot random forest regression model or use it for this use case I also processed the values of the X_test in standard scalar and label encoded all the categorical data. My prediction using the random forest regressor for data like : 'age':19,'sex':'male','bmi':27.9,'children':0,'smoker':'yes','region':'northeast' is coming out as 17207.38 which is close to the actual value
Hi brother Nice work you are doing. I have a doubt at last after running the predictive system there is a huge variation in my prediction and there is a User Warning with x does not have valid feature name. Can I know your suggestion on this?
Random Forest Regression gave me 83.47% test accuracy. BTW, please teach hyperparameter tuning and ensemble methods for each ML problem to help in improving the accuracy!
hi! this machine learning course is all about understanding Machine Learning concepts & implementing. after this, I'll make Deep Learning course. then I'll make deployment videos. that's what I have planned for now. thanks! but I'll definitely consider ur request.
@siddhardhan want to ask while building model with DTR, with feature_importance can I drop the region feature becz its playing the least role in the model
Hi, I was watching project 12 which was of sales forecasting. In that whole video I expected that there will be a chart which will show how sales will be in future. It was not there the video ended at R2 SQAURE. Than I read through comments and found out someone else had some question and you directed him to this video. Now in this video predictive system is build and video ended there. Now will you direct me to another video? I am a fresher and want to do some complete project and your video seems nice but if I have to go from 1 video to another I get confused. In short I have no idea how to complete my project on sales forecasting of big mart. Any direction will be good. Thank you😄
Hello Siddardhan bro , can this idea of predicting Medical Insurance cost would be convincing to the Insurance company as we are considering the age ? For instance I have a 18 year old student who had filled the inputs (for example) like his gender , bmi , smoker / non smoker.... If he got a certain value after giving his inputs , that value could be really reliable as he was a student still..
Hi Rahul! It is a very broad topic. I'll make a separate video on model selection in "Model Training" module. You can go through the course curriculum. I have given for course curriculum file in the video description
As of now, make note of this: Classification problems: 1. Logistic regression 2. Support vector machine 3. Random forest, etc. Regression problems: 1. Linear regression 2. Svm for regression 3. Xgb rrgressor, etc Image recognition: 1. Convolutional neural networks Speech data : 2. Recurrent neural networks
@@Siddhardhan also if possible next time can you use flask for model deployment/model predictive system and i take data as input from the user then predict the value is right in predictive system like age=int(input("enter your age")).....for 5 features then inputdata=(age,bmi...,smoker) inputarray=np.array(inpiutdata) predict=regressor.predict(inputarray.reshape(1,-1))
Hi Rahul! Making deployment videos at the moment, doesn't fit the machine learning course. The contents of this course will be centred around Machine Learning models, their math, data cleaning and other things. After this machine learning course, I'll make a deep learning course. Then we can go with deployment. That would be a better order for Learning.
hi! as there are several categorical columns in this dataset, I didn't use standard scaler. you can try it anyways and see whether it's useful in this case.
In have a small dought in the section:-Building a predictive section... What is the need of reshaping it to (1,-1).... what does that mean actually.... kindly give an explanation for this.... thank you
Hey, Is there a reason why are you encoding regions as 0,1,2,3 ? If you encode like that it means one region is greater than other. Shouldn't it be one hot encoding ?
you can practice this code well & put it in your resume. but I am not sure whether you can deploy it in some application. the dataset might be copyrighted.
while encoding the categorical variables, I'm getting this error: data.replace({'sex':{'male':0, 'female':1}}, inplace=True) --> CODE AttributeError: 'numpy.ndarray' object has no attribute 'replace' --> ERROR What should I do ?
crossplot "actual vs. prediction" shows that they are not fit very well. The plotting script: plt.scatter(Y_train, training_data_prediction) plt.xlabel("Actual") plt.ylabel("Predicted") plt.show()
![Lp. Сердце Вселенной #60 РОЖДЕНИЕ ЛОЛОЛОШКИ [Финал] • Майнкрафт](http://i.ytimg.com/vi/YoR0pAV9FVQ/mqdefault.jpg)
Thank you so so much bro. I have completed my semester project yesterday. You are the best educational UA-cam channel ever watched 💗🥺
nice profile pic.
Please help me in this project
Could you give any index for this project..plz
Thank you sir ,
To explain this project in proper way
😀
why standardization is not done in this project?
my prediction on test case goes wrong why?
Great explanation!
your explaination is best, thank u so much
Why don't you find any Outlier, and don't you find the correlation between the feature and the target ?
Amazing explanation
very good work, very well explained
Great stuff
what is the future scope of this or how is this project useful for the real life scenario
Can any tell the insurance cost in the data set is paid by insurance company or by patient
test accuracy for : simple linear regression : 0.76
svm regression: -0.11
random forest regression: 0.803
so our choice right now can be to plot random forest regression model or use it for this use case I also processed the values of the X_test in standard scalar and label encoded all the categorical data. My prediction using the random forest regressor for data like : 'age':19,'sex':'male','bmi':27.9,'children':0,'smoker':'yes','region':'northeast' is coming out as 17207.38 which is close to the actual value
Hi brother
Nice work you are doing. I have a doubt at last after running the predictive system there is a huge variation in my prediction and there is a User Warning with x does not have valid feature name. Can I know your suggestion on this?
I had the same issue. The predictions are not accurate....the predicted value of the exact input data shown in the video is only accurate.
Random Forest Regression gave me 83.47% test accuracy. BTW, please teach hyperparameter tuning and ensemble methods for each ML problem to help in improving the accuracy!
Can you provide the code
which are all the tools used in this project please confirm
Quick question, why can't we use label encoding instead of manually replacing values. I'm new to these
Great work.
Request you to make end to end project including front end development and database connectivity.
hi! this machine learning course is all about understanding Machine Learning concepts & implementing. after this, I'll make Deep Learning course. then I'll make deployment videos. that's what I have planned for now. thanks! but I'll definitely consider ur request.
can you please help me telling how to deploy machine learing in backend and html,css in frontend ? Which tutorial you followed please sir ? Help
How to create a website for this project sit
@siddhardhan want to ask while building model with DTR, with feature_importance can I drop the region feature becz its playing the least role in the model
smokers data is not converting into 1's and 0's after applying replace? can you tell the solution
Hi, I was watching project 12 which was of sales forecasting. In that whole video I expected that there will be a chart which will show how sales will be in future. It was not there the video ended at R2 SQAURE. Than I read through comments and found out someone else had some question and you directed him to this video. Now in this video predictive system is build and video ended there. Now will you direct me to another video? I am a fresher and want to do some complete project and your video seems nice but if I have to go from 1 video to another I get confused. In short I have no idea how to complete my project on sales forecasting of big mart. Any direction will be good. Thank you😄
Sir why i am getting my charges value prediction more and more than actual data
Really appreciate!👏🏻
My pleasure 😊
Thanks a lot !
Great effort. Thanks
my pleasure 😇
Sir my linear regression is not executing why sir
Does preprocessing done here and missing value? Any more things we can include in this?
hi! sometimes we may have to handle the outliers.
Thank you
Welcome 😇
Hello Siddardhan bro , can this idea of predicting Medical Insurance cost would be convincing to the Insurance company as we are considering the age ? For instance I have a 18 year old student who had filled the inputs (for example) like his gender , bmi , smoker / non smoker.... If he got a certain value after giving his inputs , that value could be really reliable as he was a student still..
wonderful
Is this project is similar to customer lifetime value about insurance
Thanks alot sir
Well explained
thanks a lot 😇
How to create te endpoint ?
how we choose which model to use for prediction?
Hi Rahul! It is a very broad topic. I'll make a separate video on model selection in "Model Training" module. You can go through the course curriculum. I have given for course curriculum file in the video description
As of now, make note of this:
Classification problems:
1. Logistic regression
2. Support vector machine
3. Random forest, etc.
Regression problems:
1. Linear regression
2. Svm for regression
3. Xgb rrgressor, etc
Image recognition:
1. Convolutional neural networks
Speech data :
2. Recurrent neural networks
@@Siddhardhan also if possible next time can you use flask for model deployment/model predictive system and i take data as input from the user then predict the value is right in predictive system
like age=int(input("enter your age")).....for 5 features
then inputdata=(age,bmi...,smoker)
inputarray=np.array(inpiutdata)
predict=regressor.predict(inputarray.reshape(1,-1))
Hi Rahul! Making deployment videos at the moment, doesn't fit the machine learning course. The contents of this course will be centred around Machine Learning models, their math, data cleaning and other things. After this machine learning course, I'll make a deep learning course. Then we can go with deployment. That would be a better order for Learning.
@@Siddhardhan oka and thanks really looking forward to this
Hi , as there is scale variation in the fields.Shouldn' we go for Standard Scaler before fitting. Can you please advise.
hi! as there are several categorical columns in this dataset, I didn't use standard scaler. you can try it anyways and see whether it's useful in this case.
hi, how can you calculate the P value?
Awesome video, keep up the incredible work! :)
Thanks a lot 😇
In have a small dought in the section:-Building a predictive section...
What is the need of reshaping it to (1,-1).... what does that mean actually.... kindly give an explanation for this.... thank you
bro can we work on jupyter notebooks ?? will it be simmilar to it
Yes, you can. But please practice with pycharm & vscode as well. Notebooks are good only for exploratory purpose.
hi sir if I want using Genetic algorithm what can I do for it if you explain this algorithm before pleas send me video
thanks a lot to you
Hello sir in this project put in my resume but what will be the description put the resume please Sujjection
Hey, Is there a reason why are you encoding regions as 0,1,2,3 ? If you encode like that it means one region is greater than other. Shouldn't it be one hot encoding ?
4 classes does not cause any problem. you can definitely try one hot encoding as well. there's nothing wrong with it.
I need this project. How i will get?
Sir why u are not scaling the data?
30:55
The mean absolute error is around 2000 .
How can it be improved
ربنا يسترك يا جدع روح
Help me with this error
Value error: could not convert string to float: 'northeast'
hi! check your label Encoding part. there may be a minor mistake.
Hi sir I want to learn in private from u what is the procedure to apply
Hi! I am not giving one-to-one training as of now.
May I put this project in my resume as a fresher if I also convert this project into the flask and deploy it on Heroku?
you can practice this code well & put it in your resume. but I am not sure whether you can deploy it in some application. the dataset might be copyrighted.
@@Siddhardhan sahi sahi, thanks
لماذا استخدمنا خوارزمية linerTegession
Display screen looks blurr.. Not properly visible.
can i put this on my resume ? Is this project count ?
No
9:00
can I know the accuracy of this prediction?
we don't calculate accuracy for Regression problems. there are other values like r squared value, mean absolute error, etc. please refer the video.
@@Siddhardhan actually if I tried for other inputs which are there in dataset I'm not getting at least near values like I'm getting double of actual
Try optimising the model
@@vaishnaviyada even I'm facing the same problem
@@Siddhardhan how am I supposed to do that?
bro some of the results are varying with a very big diffrence
Allam.o.Alaikum sir
Hello sir!can you send the ppt
NameError: name 'regressor' is not defined ??
You may not have imported it
from sklearn.linear_model import LinearRegression
Run this
@@Siddhardhan got it
Great 😇
same isssue happens me also
Great work! I have a question, why do we always use this dataset? Do insurance companies still use this data today? Can we add features?
hi! it's not like that. this dataset is for demonstration. companies will have their own data.
@@Siddhardhan Yep! Thanks very much! can you suggest me some more features to improve the data?
Like?
while encoding the categorical variables, I'm getting this error:
data.replace({'sex':{'male':0, 'female':1}}, inplace=True) --> CODE
AttributeError: 'numpy.ndarray' object has no attribute 'replace' --> ERROR
What should I do ?
try restarting the runtime and run the cells again. check the code given in description
the model is totally wromg if taken any other input raather then you used your model gives vague solution
I need your in python plz help me
write a mail to datascience2323@gmail.com
Lekin accuracy tih 77% hai
i dont understand any of this :DDDDDDDDDDDDDDDDDDDDDDDDDDD
crossplot "actual vs. prediction" shows that they are not fit very well. The plotting script:
plt.scatter(Y_train, training_data_prediction)
plt.xlabel("Actual")
plt.ylabel("Predicted")
plt.show()