This is my first attempt to study AI. Recently I achived collecting data energy for large buildings without any software installed in the property. This tutorial was extremely helpful to understand the basics of predictons. I will love a second part going deep into the last section of this video, i felt a little bit lost but is my lack of experience in this subjects. Kudos Sir!
Nice tutorial but I don't understand why almost everyone just do prediction and compare to the actual data. But this is not necessary without forecasting the future values. Forecasting future values are the most important not just predictions
Hello, I still do not understand how to predict future values. This video shows how to test the model using existing data to create X_test. How do I predict next week's values if I have no data_test set, because next week's data_test values are unknown?
Curtis Morgan hi, use dataset till today's to predict tomorrows value and then include tomorrows value in dataset to predict day after tomorrow's value and so on.
@@LaxmiKantTiwarilaxmimerit But how can I do that when our model outputs one value, and our data set has five different columns of data? To which column do I add the new value?
Curtis Morgan I my problem is maybe even more complicated, because I want to use also information like outside temperature to have better score. But if you only use power consumption, I think that this paper might be helpful machinelearningmastery.com/how-to-develop-lstm-models-for-multi-step-time-series-forecasting-of-household-power-consumption/
If you have a look at the description from the repository, the data is from France and the energy consumption is higher during winter time and low during the summer.
Hi, Great video thank you. I have a question on a specific case: I want to use LSTM with a simple NN for regression purpose (yield prediction) that use the biomass of a plant across his growing period. I want to know if the addition of date increase the performances and therefore dynamically predict this yield for the growing season. Do you think it possible?
Amazing video !! Thank you very much . I have a question ..In order to forecast properly and to evaluate the moel , we should update the previous values with the predicted ones in the test phase . Shouldn't we ?
Thank you for the excellent tutorial - I presume if you use all columns then we have to change the input shape - however what if we want to predict let's say for future 1 year - would that need 365 dense layers?
Hi, Very nice tutorials. I have a question. I have some data with 8 features (including the target variable to predict which is Demand). My question is how to do a multistep step prediction for only one variable (for this case, Demand) using all the features from the previous time steps. I guess I would have to include all the columns in the line 66 of this file x_train.append and only one column in the y_train.append, but I am not sure if that's the correct way. Thanks and keep up the great work!
I have a similar kind of dataset which is based on landslides, but it is like classification problem. For such problems how to calculate the accuracy of the model?
def fill_missing(data): one_day = 24*60 for row in range(data.shape[0]): for col in range(data.shape[1]): if np.isnan(data[row, col]): data[row, col] = data[row-one_day, col] Sir, in this code you've written row-one_day but you said that the missing value is filled by the previous year value of the same day? What is this sir? Please explain
Hi, This is in regard to the code files of all videos at KGP Talkie UA-cam Channel. I request you to please get enrolled yourself in my Udemy course and show your support and love to KGP Talkie. All the code files and video lectures have lifetime access with 30 Days money back Guarantee. Code and question-answer support are also available at Udemy. I promise you to give FREE COUPONS for the next course on Deep Learning and ML. You can click on the link mentioned below and can get yourself enrolled!! bit.ly/udemy95off_kgptalkie New content is added at Udemy: 1. Animation Plot [2 lectures] 2. Python Coding in Mobile [5 lectures] 3. Complete EDA of Boston Dataset [20 lectures] What else we promise in this course 1. Kaggle data EDA 2. Text data EDA 3. More Animation Plot 4. More 3D plots 5. Figure Aesthetics and Decoration 6. Free coupons for next course 7. And so much more. Hurry up!!! Only for a limited time. Please email your details at udemy@kgptalkie.com for the FREE COUPONS of the next course. Note: Code files of UA-cam lectures will be also available once you register in this course. Please send an email to udemy@kgptalkie.com after enrolling in this course.
Thanks for posting this, it really helped me get through an issue I couldnt understand. You have a great way of distilling things down and demonstrating them effectively as you go. Any chance you could update the github link? The one provided gives a 404.
I think you made some mistake adding the power. You can add energy but power I donot think. Like temperature, you can not add the monthly temperature and say in a year we have this sum of temperature in a particular location, instead you take average.
what is the point of doing so much of EDA on every column when you were going to take only one column only in the training.. Didnt like the approach.. Sorry...
This is my first attempt to study AI. Recently I achived collecting data energy for large buildings without any software installed in the property. This tutorial was extremely helpful to understand the basics of predictons. I will love a second part going deep into the last section of this video, i felt a little bit lost but is my lack of experience in this subjects. Kudos Sir!
super helpful and i actually love ur voice :}
Thank you so much..!!! one of the bestest detailed explanation available !!!
Nice tutorial but I don't understand why almost everyone just do prediction and compare to the actual data. But this is not necessary without forecasting the future values. Forecasting future values are the most important not just predictions
exactly......
If we dont compare to exisitng data, how will we verify the accuracy of the developed algorithm 🤔
What a great tutorial!!!
Thank you for all your efforts and dedication.
Thank you so much for watching ❤️
thank you very much!!!!! Really easy to follow.
Good Job!!! Thanks for sharing!!!
Maam, your explanation is better than what I learned in my class.
Well illustrated! Thanks
Maaaaan !!!!! You are a GEM. Such an amazing explanation
Thank you so much
Thanks for the wonderful work done. I really enjoyed the demo of your lecture. It has actually helped me a lot.
you are the best
These are very good tutorials, please keep this kind of work. U are the best!!
Thank you so much 😍😍
Such a great tutorial. Helped me a lot. Thanks for the time
Wow, loved the explanation.
Thank you so much for watching. Please keep watching and please do share it with your friends.
sir ,
great video and your explanation is awesome.
Please sir make one video on multivariate Lstm using weather dataset.
Can you show how to plot and predict for any given future dates.?
Hello, I still do not understand how to predict future values. This video shows how to test the model using existing data to create X_test. How do I predict next week's values if I have no data_test set, because next week's data_test values are unknown?
Curtis Morgan hi, use dataset till today's to predict tomorrows value and then include tomorrows value in dataset to predict day after tomorrow's value and so on.
@@LaxmiKantTiwarilaxmimerit But how can I do that when our model outputs one value, and our data set has five different columns of data? To which column do I add the new value?
@@thecurtismorgan Did you find answer? I have the same problem
@@bargzo I'm not sure, but I think we need to set the model up a multivariate AND multistep model. I have not been able to do this yet.
Curtis Morgan I my problem is maybe even more complicated, because I want to use also information like outside temperature to have better score. But if you only use power consumption, I think that this paper might be helpful machinelearningmastery.com/how-to-develop-lstm-models-for-multi-step-time-series-forecasting-of-household-power-consumption/
Hello !! please help your Github Link is showing error for the last 20 days
still now it is showing error
Excellent this series..
Thank you so much for watching.
If you have a look at the description from the repository, the data is from France and the energy consumption is higher during winter time and low during the summer.
Hi, Great video thank you. I have a question on a specific case: I want to use LSTM with a simple NN for regression purpose (yield prediction) that use the biomass of a plant across his growing period. I want to know if the addition of date increase the performances and therefore dynamically predict this yield for the growing season. Do you think it possible?
Amazing video !! Thank you very much .
I have a question ..In order to forecast properly and to evaluate the moel , we should update the previous values with the predicted ones in the test phase . Shouldn't we ?
yes. we can do that by using online sequential LSTM model to feed testing examples one by one into it.
Why is your Github Link not working ? Its showing 404 error
Did you use walk forward approach?
@KGP Talkie is there a way to time series predict multiple columns at one time.
Also, it may help if you can plot for future 6 months if possible and show with past data! Thanks again
Very nice, man. In the end, you could have plotted the prediction vs the true values and the train loss graph aswell
Did you find how to do that ?
Thank you for your great explanation.
Do you have any plan to explain multi-variate LSTM?
can help me to understand how to implement GRU for prediction??
Thank you for the excellent tutorial - I presume if you use all columns then we have to change the input shape - however what if we want to predict let's say for future 1 year - would that need 365 dense layers?
Hi,
Very nice tutorials. I have a question. I have some data with 8 features (including the target variable to predict which is Demand). My question is how to do a multistep step prediction for only one variable (for this case, Demand) using all the features from the previous time steps. I guess I would have to include all the columns in the line 66 of this file x_train.append and only one column in the y_train.append, but I am not sure if that's the correct way.
Thanks and keep up the great work!
Which type of LSTM is this for Multi Step Ahead Prediction?
Why do you only make one lstm layer for this regressor?
thx a lot.
I have a similar kind of dataset which is based on landslides, but it is like classification problem. For such problems how to calculate the accuracy of the model?
def fill_missing(data):
one_day = 24*60
for row in range(data.shape[0]):
for col in range(data.shape[1]):
if np.isnan(data[row, col]):
data[row, col] = data[row-one_day, col]
Sir, in this code you've written row-one_day but you said that the missing value is filled by the previous year value of the same day? What is this sir? Please explain
Sir please do a video on building a movie recommendation system with deployment
Yeah sure. I am working on it.
Is it matter Multistep Prediction or Multistep Ahead Prediction? Or both are having same meaning?
How do I get this code, please? This is the BEST.
Thanks a lot sir
can anyone please elaborate what is y_true?
Y true is true value of the data
Hi,
Can you explain a little more on how to do a multi step prediction with multiple features using LSTM. Thanks!
Hi,
This is in regard to the code files of all videos at KGP Talkie UA-cam Channel. I request you to please get enrolled yourself in my Udemy course and show your support and love to KGP Talkie. All the code files and video lectures have lifetime access with 30 Days money back Guarantee. Code and question-answer support are also available at Udemy.
I promise you to give FREE COUPONS for the next course on Deep Learning and ML. You can click on the link mentioned below and can get yourself enrolled!! bit.ly/udemy95off_kgptalkie
New content is added at Udemy:
1. Animation Plot [2 lectures]
2. Python Coding in Mobile [5 lectures]
3. Complete EDA of Boston Dataset [20 lectures]
What else we promise in this course
1. Kaggle data EDA
2. Text data EDA
3. More Animation Plot
4. More 3D plots
5. Figure Aesthetics and Decoration
6. Free coupons for next course
7. And so much more.
Hurry up!!! Only for a limited time.
Please email your details at udemy@kgptalkie.com for the FREE COUPONS of the next course.
Note: Code files of UA-cam lectures will be also available once you register in this course. Please send an email to udemy@kgptalkie.com after enrolling in this course.
If i Want to change the predict intervel, example 1 week to 1 month, how i do that?
SIR PLEASE UPLOAD ARTIFICIAL NEURAL NETWORK LECTURE WITH TF BY USING CSV DATASET
Yeah sure. Will try soon.
Thanks for posting this, it really helped me get through an issue I couldnt understand. You have a great way of distilling things down and demonstrating them effectively as you go. Any chance you could update the github link? The one provided gives a 404.
Hi,
Codes are being uploaded at kgptalkie.com. Deep Learning codes will be up soon.
Can anybody tell me how to read the prediction? Where is the first day, 2nd day until 7th day future? And how to make plot for the prediction?
Make video on cnn lstm on colour image
Sir, Please do some videos on GANs.
Yeah Sure Sai... I will try it soon.
I think you made some mistake adding the power. You can add energy but power I donot think. Like temperature, you can not add the monthly temperature and say in a year we have this sum of temperature in a particular location, instead you take average.
Sir post some video on this library
Yeah sure. I will make more on LSTM.
sir font size is small
It's an old video. Older videos had small font.
what is the point of doing so much of EDA on every column when you were going to take only one column only in the training.. Didnt like the approach.. Sorry...