Today, I will talk about some features of the upcoming version of SPMF 2.63 that I am working on and is planned for release next month.
Improved Pattern Visualizer window
The first feature is an improvement of the Pattern Viewer tool so that it can display the richer SPMF format where items have names (strings) in patterns. Previously, a pattern file like this:
@ITEM=1=apple
@ITEM=2=orange
@ITEM=3=tomato
@ITEM=4=milk
@ITEM=5=bread
1 2 4 5
2 3 5
1 2 4 5
1 2 3 5
1 2 3 4 5
2 3 4
would be displayed as this by the Pattern Viewer:
Now, it is displayed as:
Other pattern visualizations
I am also working on adding other tools for visualization such as a Rule Viewer to visualize association rules and sequential rules like this:
The above screenshot is an early version. I will improve the appearance of this window and I still have to think about how to best integrate it in the software.
Today, I just wanted to show you some ideas of new features. If you have any ideas or comments, send me an e-mail or leave a comment below!
Hello sir,
As many methods are building tree structures and use search function on those constructed trees for HUP & CoHUP mining, it will be very helpful to see how a tree is build and pruned at each insert operation on varying minUtil or minCorr values.
Also, if we can have a side by side windows/tabs to compare performances of algorithms and check same and different patterns that are getting mined, then it can really help in preparing observations of research.
I hope the suggestions are relevant and I am excited to see it in the pattern viewer tool.
Reference:
https://data-mining.philippe-fournier-viger.com/upcoming-features-for-spmf-2-63/
Thanks for your suggestions!
Hello sir,
As many methods are building tree structures and use search function on those constructed trees for HUP & CoHUP mining, it will be very helpful to see how a tree is build and pruned at each insert operation on varying minUtil or minCorr values.
Also, if we can have a side by side windows/tabs to compare performances of algorithms and check same and different patterns that are getting mined, then it can really help in preparing observations of research.
I hope the suggestions are relevant and I am excited to see it in the pattern viewer tool.
Reference:
https://data-mining.philippe-fournier-viger.com/upcoming-features-for-spmf-2-63/