Thank you so much, Philippe. This video is so helpful. I am wondering if you ever plan to do a demonstration using R Studio. I also do not mind if you can point me to any material that does that. Thanks
Hi, thanks for your nice message. I am not very familiar with R. I am not sure if it can do sequential pattern mining. But Rcan be used for association rule mining (which is a bit different). I have a copy of this book in my office: Zhao, Y. (2012) R and Data Mining: Examples and Case Studies, Academic Press, Elsevier. It is an old book but it has a chapter that explains how to use R for association rule mining with some examples. By the way, if you want to call SPMF (which is in Java from R), I found that there is a wrapper in Github for some algorithms. Maybe it can allow to call sequential pattern mining algorithms from R: github.com/pommedeterresautee/spmf/tree/master/RCaller
Great video, I'm doing my final master thesis about finding the most freqüent patterns in overweight babies using their data along parents data. I will use the SPADE algorithm.
@@philfv Dear Philippe, if I have 400 babies (instances) and 5 of the features are the weight of the baby at birth, month2, month4,month6 and first year. How could I discretize this 5 features? I guess that SAX will not work well. Im thinking of calculate the quartiles of each feature and assign a label to each feature depending on the quartile that belongs to.
@@raizen74 Good evening, I see. 1) For SAX, yes, it is for time series but i think the idea of SAX could be adapted for discretizing a single numeric attribute. I think it could work but you would still have to choose the number of bins (labels) for the discretization. And SAX has the assumption that the distribution is a normal distribution. But it could be adapted to other distribution. 2) The idea of using quartile could be also a solution. 3) There exists other discretization methods as well. I am not familiar with all of them. But I think they would have different advantages/disadvantages. So maybe you should first check the distribution of your data to see how it looks like and this could maybe guide which method to use?
Thanks for your appreciation and watching🙂 I am currently teaching at Shenzhen University in the south of China. If you like the channel and want to support it, please subscribe ;-)
Yes, sure, it is on that page with many other related PPTs for pattern mining: www.philippe-fournier-viger.com/COURSES/Pattern_mining/index.php And that PPT in particular is here: www.philippe-fournier-viger.com/COURSES/Pattern_mining/Sequential_pattern_mining.pptx
If you want more pattern mining videos, check my channel ;-)
Hey; thank you for such an amazing tutorial. The combinations of topics alongside the "this means" in your videos have helped me greatly.
Thanks for the feedback! Happy it is useful
Thanks for the video, helped me a lot in starting this topic!
Thank you very much. Very well explained sequence algorithm
Amazing expalantion. Short and clear.
It would be even better if you could add links to videos about the algorithms mentioned...
Thanks for your nice comments and the feedback. I will take this into account in the future and try to add links between videos.
Thank you so much, Philippe. This video is so helpful. I am wondering if you ever plan to do a demonstration using R Studio. I also do not mind if you can point me to any material that does that. Thanks
Hi, thanks for your nice message. I am not very familiar with R. I am not sure if it can do sequential pattern mining. But Rcan be used for association rule mining (which is a bit different). I have a copy of this book in my office: Zhao, Y. (2012) R and Data Mining: Examples and Case Studies, Academic Press, Elsevier. It is an old book but it has a chapter that explains how to use R for association rule mining with some examples. By the way, if you want to call SPMF (which is in Java from R), I found that there is a wrapper in Github for some algorithms. Maybe it can allow to call sequential pattern mining algorithms from R: github.com/pommedeterresautee/spmf/tree/master/RCaller
very good and clear explanation sir, thanks for making this video
Thanks for watching! Happy it helps
Great video, I'm doing my final master thesis about finding the most freqüent patterns in overweight babies using their data along parents data. I will use the SPADE algorithm.
Thanks. Nice topic. Wish you good results 🙂
@@philfv Dear Philippe, if I have 400 babies (instances) and 5 of the features are the weight of the baby at birth, month2, month4,month6 and first year. How could I discretize this 5 features? I guess that SAX will not work well. Im thinking of calculate the quartiles of each feature and assign a label to each feature depending on the quartile that belongs to.
@@raizen74 Good evening, I see. 1) For SAX, yes, it is for time series but i think the idea of SAX could be adapted for discretizing a single numeric attribute. I think it could work but you would still have to choose the number of bins (labels) for the discretization. And SAX has the assumption that the distribution is a normal distribution. But it could be adapted to other distribution. 2) The idea of using quartile could be also a solution. 3) There exists other discretization methods as well. I am not familiar with all of them. But I think they would have different advantages/disadvantages. So maybe you should first check the distribution of your data to see how it looks like and this could maybe guide which method to use?
Thank you, Sir! Great explanatory video. Any plan to add a video about Colossal Pattern Mining?
Thanks! At this moment, no. But I will keep this as a suggestion for a future video. Thanks for this suggestion.
Pr., where do you teach? You are great!
Thanks for your appreciation and watching🙂 I am currently teaching at Shenzhen University in the south of China. If you like the channel and want to support it, please subscribe ;-)
Thank you so much Sir! Great video!
very beneficial video, thank you sir
Thanks for watching!
Do you provide 1-on-1 coaching for dissertation?
Sorry, no.
@@philfvok. Ty for your reply anyways
can you share this great ppt?
Yes, sure, it is on that page with many other related PPTs for pattern mining:
www.philippe-fournier-viger.com/COURSES/Pattern_mining/index.php
And that PPT in particular is here:
www.philippe-fournier-viger.com/COURSES/Pattern_mining/Sequential_pattern_mining.pptx
@@philfv Thank you very much for your sharing
thanks