You are the best explainer in this topic i've found in UA-cam. The indian guys i've found are really amazing, but the bad english and the lack of focus are barriers to acknowledgement. Congratulations and thank you for have a script before recording the video and for being so patient during it.
Video of useful but I doubt interpretation of variance explained by model. shouldn't we be looking at proportion and cumulative columns for variance explained rather than looking at ranked attributes figure?
Hi, regarding feature extraction using PCA I applied PCA filter on input dataset that consists of 13 predictors and got 18 principal components. Is it possible? At first, the input dataset consists of 6 missing values with that it produced 12 PCs when replace the missing values with majority mark of that particular features I got 18 PCs. May I know what exactly the concept behind this? By the way I used Cleveland heart dataset and WEKA platform. Thank you.
Great explanation sir. but what after that ?? how to use this data in my model. I'm making a IDS model for detecting attacks and I need to reduce the features to get faster and accurate detection
Great content. Thank you so much for explaining it. But I think it's worth pointing out that the ranked values don't add up to 1. When you try with other datasets, these values are all less than 1, but certainly aren't adding up to 1. For example, I got 0.74, 0.52, 0.39 for the first 3 attributes (there were more) for the diabetes dataset. I would say that it was just a coincidence the values got in the video, but even in the cpu dataset that was used in the video, they don't add to 1 as well. They appear to, at a first look, but when you actually add them up, you get 1.0473. Close, but not 1. Thanks anyway, all the rest is really helpful.
Hello. Really practical video. I have one question: How do you define the threshold? I mean why 50% or 40%? Do these numbers come out of random? During the first phase, I realized that it is easy to select features since they are ranked ( so they use something like entropy algorithm or so). But what about these percentages that you are mentioning? Where do they come from? Thank you in advance!
You are the best explainer in this topic i've found in UA-cam. The indian guys i've found are really amazing, but the bad english and the lack of focus are barriers to acknowledgement. Congratulations and thank you for have a script before recording the video and for being so patient during it.
Thank you very much. I found this tutorial is the bestest for pca in weka. Thank you very much once again.
This is such a great explanation. Thank you
For some reason my ranked attribute variance doesn't add up to. Does anyone know whats the isssue with it
Hozaifah Zafar I got the same problem
i think the one that add up to 1 is the proportion part of the PC, not the one he pointed out
Hi Chris great video can I ask a question please
Video of useful but I doubt interpretation of variance explained by model. shouldn't we be looking at proportion and cumulative columns for variance explained rather than looking at ranked attributes figure?
Hi, regarding feature extraction using PCA
I applied PCA filter on input dataset that consists of 13 predictors and got 18 principal components. Is it possible?
At first, the input dataset consists of 6 missing values with that it produced 12 PCs when replace the missing values with majority mark of that particular features I got 18 PCs. May I know what exactly the concept behind this?
By the way I used Cleveland heart dataset and WEKA platform.
Thank you.
Great explanation sir.
but what after that ?? how to use this data in my model. I'm making a IDS model for detecting attacks and I need to reduce the features to get faster and accurate detection
Great content. Thank you so much for explaining it.
But I think it's worth pointing out that the ranked values don't add up to 1. When you try with other datasets, these values are all less than 1, but certainly aren't adding up to 1. For example, I got 0.74, 0.52, 0.39 for the first 3 attributes (there were more) for the diabetes dataset. I would say that it was just a coincidence the values got in the video, but even in the cpu dataset that was used in the video, they don't add to 1 as well. They appear to, at a first look, but when you actually add them up, you get 1.0473. Close, but not 1.
Thanks anyway, all the rest is really helpful.
Hello. Really practical video. I have one question: How do you define the threshold? I mean why 50% or 40%? Do these numbers come out of random? During the first phase, I realized that it is easy to select features since they are ranked ( so they use something like entropy algorithm or so). But what about these percentages that you are mentioning? Where do they come from? Thank you in advance!
Awesome job with this tutorial, very helpful thanks
Hello, quick question. Should the data be normalized before I use PCA? Thanks for sharing.
thanks for sharing, very good tutorial!
hey thanks a lot, the tutorial really helpful !!
Can I use PCA with dataset containing non-numeric attributes?
no
please could you do SVM and J48 in weka same like this with clear explanation soon as you can. Thank you
My ranked attribute values aren't adding upto 100.. if anyone could explain what am I doing wrong?
6:25