+ for explaining the values and formulas with Excel - for wasting time with .csv file generation - forgot to connect the data source and select column block and wondering why columns cannot be selected from a list
Thanks! This is an old video that I made when I was brand new to this stuff. I've just never taken down because it gets picked up in searches a lot. So I updated the description with a link to a full playlist that has fewer small mistakes like these
6:30 you didn't relate the imported dataset to the Select Columns part. So that's why it doesn't recognise the columns you wanted to select. (I have never used MS Azure ML before, but I think this is the cause of that.)
Unable to download data file from the link above: " This domain name expired on 11/18/17 and is pending renewal or deletion". Any other way to get the file?
This one shows you how to build the regression model in Excel: ua-cam.com/video/0Jg9vrfOYG0/v-deo.html and this one shows you how to create the calculator form: ua-cam.com/video/trPrQwt3GfM/v-deo.html
Annie Lei I need to edit my previous response. Azure ML does give you AUC, I’ve just never used it. All you have to do is visualize an "Evaluate model" pill from a classification model. AUC is in there. Both overall AUC and the cumulative AUC based on different percentages of the curve. Let me know if you have any trouble finding it. But it has to be a classification model; not a regression model like the one I use in this video. Here's an example: ua-cam.com/video/ZmVrw1oIkdg/v-deo.html
Thank you, I was able to find the AUC. I am actually doing my master in CIS. But your videos teach me so much more haha. I am just struggling with all the modules because my professor does not explain in details like how you did in your video. I was doing research, watching videos, and reading Microsoft docs all day and understood way more than learning in class. I had a task to improve the AUC score. I tried really hard to get it as high as possible but my professor wants more. If you don't mind, could you take a look at my publish and give me some comment? gallery.cortanaintelligence.com/Experiment/CIS5560-Midterm-Spr-2018-by-Hai-Anh-Le-2 My email is janele94@gmail.com. Thank you so much!
Sajjad Sheriff Thank you! And yes, this one (ua-cam.com/video/-Lf-zS7l8HI/v-deo.html) covers the one-v-all multiclass which is my favorite ensemble algorithm because it is a good combination of accuracy and minimal over fitting. If you check out my azure playlist, you will find a video for every two class classification model as well. The multi class versions of each work just the same except, obviously, that they handle more than two values for a categorical dependent variable.
Thank you mark that was quick response. Yes i have gone through two class and binary classification. But was more interested in multiclass i.e. more than two variables. Your videos are awesome.
Sajjad Sheriff cool, so the multi class versions are basically identical to the two-class except for one difference in the evaluation results. Multi class models will give you an overall accuracy score in addition to regular accuracy. Overall is like regular accuracy in a two-class model. Accuracy in a multi class model is actually the averaged accuracy of each individual categorical value predicted
Hi Marcelo, sorry about that. Somehow I missed your comment a long time ago. But anyway, the data file link should work now. But if you've been learning this stuff for the last 6 months, then you're probably well beyond this by now. Good luck! Mark
I was truly lost for a couple of days. Because of your video, tonight I feel much better. Thank you so much.
prediction using regression explained really well with an excel equivalent. Great job helped me understand the basics quickly.
Great explanation especially comparing the results from Excel to Azure. Really helps one understand between the two.
I want this data file but can't found it in link if you have plz send me on nisarali009.NA@gmail.com
Mark - this is a good comparison between excel and ML and makes me relate complex cloud branded names to the reality..Thanks , appreciate this video
Thanks!
Its good practice to always link the new module to the flow of data before making any changes.
Thank you!
Thanks a lot for the demo and the clear and helpfull explanations!
You have to connect the line from the dataset to the column selector FIRST. Then it will show you the columns.
Just a note that if you connect the Select Columns module to the dataset first, the column names should display automatically.
Thanks! Someone else also pointed that out a while back and I've just been too lazy to update the video
Thanks in a million. Great content. Awesome. Grade: A++💥
Thank you so much Mark. Take care.
Really amazing video!
This was so helpful. Thank you!
Why did you not normalize the data by calculating z-scores?
Good question, it's because Azure ML Studio automatically normalizes using the MinMax (0 to 1) approach
+ for explaining the values and formulas with Excel
- for wasting time with .csv file generation
- forgot to connect the data source and select column block and wondering why columns cannot be selected from a list
Thanks! This is an old video that I made when I was brand new to this stuff. I've just never taken down because it gets picked up in searches a lot. So I updated the description with a link to a full playlist that has fewer small mistakes like these
dont take video down,its very helping
6:30 you didn't relate the imported dataset to the Select Columns part. So that's why it doesn't recognise the columns you wanted to select. (I have never used MS Azure ML before, but I think this is the cause of that.)
Thanks - this worked!
I need to check these comments more often; thank you for pointing that out!
I was going to comment same. Nice catch 👌
Unable to download data file from the link above: " This domain name expired on 11/18/17 and is pending renewal or deletion".
Any other way to get the file?
Thanks for letting me know; I just updated that link
please i cant find the first video of how you calculated it using excel sheet...how do i find it?
This one shows you how to build the regression model in Excel: ua-cam.com/video/0Jg9vrfOYG0/v-deo.html and this one shows you how to create the calculator form: ua-cam.com/video/trPrQwt3GfM/v-deo.html
Thanks
Hi Mark, Do you know how to address the AUC in Azure ML studio?
Annie Lei I need to edit my previous response. Azure ML does give you AUC, I’ve just never used it. All you have to do is visualize an "Evaluate model" pill from a classification model. AUC is in there. Both overall AUC and the cumulative AUC based on different percentages of the curve. Let me know if you have any trouble finding it. But it has to be a classification model; not a regression model like the one I use in this video. Here's an example: ua-cam.com/video/ZmVrw1oIkdg/v-deo.html
Thank you, I was able to find the AUC. I am actually doing my master in CIS. But your videos teach me so much more haha. I am just struggling with all the modules because my professor does not explain in details like how you did in your video. I was doing research, watching videos, and reading Microsoft docs all day and understood way more than learning in class. I had a task to improve the AUC score. I tried really hard to get it as high as possible but my professor wants more. If you don't mind, could you take a look at my publish and give me some comment? gallery.cortanaintelligence.com/Experiment/CIS5560-Midterm-Spr-2018-by-Hai-Anh-Le-2 My email is janele94@gmail.com. Thank you so much!
can it predict stocks price?
sir can you plz give me the link of the dataset bcz i am not able to open it fron the given link
I want this data set excel file how i can get this plz guide...
Can you please put a link of the dataset in description?
Here you go: gamoda.org/is201/InClassDataFiles/BI/BikeBuyers.xlsx
Thank you Mark this is very useful.
Do you have any tutorials on multiclassification
Sajjad Sheriff Thank you! And yes, this one (ua-cam.com/video/-Lf-zS7l8HI/v-deo.html) covers the one-v-all multiclass which is my favorite ensemble algorithm because it is a good combination of accuracy and minimal over fitting. If you check out my azure playlist, you will find a video for every two class classification model as well. The multi class versions of each work just the same except, obviously, that they handle more than two values for a categorical dependent variable.
Thank you mark that was quick response. Yes i have gone through two class and binary classification. But was more interested in multiclass i.e. more than two variables. Your videos are awesome.
Sajjad Sheriff cool, so the multi class versions are basically identical to the two-class except for one difference in the evaluation results. Multi class models will give you an overall accuracy score in addition to regular accuracy. Overall is like regular accuracy in a two-class model. Accuracy in a multi class model is actually the averaged accuracy of each individual categorical value predicted
we need time series models video
I agree; it’s on my “to-do list” for next next semester. I keep putting it off ;)
Some day soon
It was better to use Summarize Data
link broken, sir
update the dataset please
link
Hi Marcelo, sorry about that. Somehow I missed your comment a long time ago. But anyway, the data file link should work now. But if you've been learning this stuff for the last 6 months, then you're probably well beyond this by now. Good luck! Mark