create FP-Growth graph
TobiasNehrig
MemberPosts:41
Guru
Hi Experts,
I've a questions about creating a graph form the results of the FP-Growth operator without using the Create Association Rules operator. Is there a way to visualize the FP-Growth results in a graph?
<运营商激活= " true "类=“子流程”薪酬atibility="8.2.001" expanded="true" height="82" name="Crawler" width="90" x="45" y="34">
<运营商激活= " true "类=“子流程”薪酬atibility="8.2.001" expanded="true" height="82" name="Crawler Spon" width="90" x="45" y="34">http://www.spiegel.de"/>
<运营商激活= " true " class = " free_memory“compatibility="8.2.001" expanded="true" height="82" name="Free Memory (8)" width="90" x="246" y="34"/>
<运营商激活= " true " class = " free_memory“compatibility="8.2.001" expanded="true" height="82" name="Free Memory (7)" width="90" x="648" y="34"/>
<运营商激活= " true " class = " free_memory“compatibility="8.2.001" expanded="true" height="82" name="Free Memory (9)" width="90" x="380" y="34"/>
<运营商激活= " true "类=“子流程”薪酬atibility="8.2.001" expanded="true" height="124" name="Prepare Data" width="90" x="246" y="34">
<运营商激活= " true "类=“子流程”薪酬atibility="8.2.001" expanded="true" height="82" name="fp Growth" width="90" x="514" y="34">
<运营商激活= " true " class = " free_memory“compatibility="8.2.001" expanded="true" height="82" name="Free Memory (4)" width="90" x="849" y="34"/>
<运营商激活= " true "类=“子流程”薪酬atibility="8.2.001" expanded="true" height="82" name="Co-occurrence" width="90" x="514" y="34">
<连接from_port = " 1 " to_op = "过程文档from Data" to_port="example set"/>
best regards
Tobias
Tagged:
0
Best Answer
-
TobiasNehrig
MemberPosts:41Guru
Hi,
i've found a solution to create a co-occurrence graph based the approach of@bhupendra_patil. After writing the FP-Growth result in a XML-File, I had to read the XML-File two times and create a new ExampleSet.
<参数键= value =“use_default_namespace歧视e"/>
<参数键= value =“use_default_namespace歧视e"/>
<连接from_op = "重命名Word1”from_port = "的例子set output" to_op="Join (2)" to_port="right"/>Tobais
1
Answers
Not that I know of, but I would be interested if any other community members know a way to do this!
Lindon Ventures
Data Science Consulting from Certified RapidMiner Experts
Hi@Telcontar120,
I found this post "Writing Association Rules to Exampleset or file" from@bhupendra_patiland I've tried to implement this in my process. But writing the FP-Growth result in a XML file blows nearly my RAM (32 GB) and creates a 8GB file. The mentioned Read XML Operation blows finally my RAM and the Process terminates.
I'm curious as to which version of RM Studio you are using. 8.1 and below has the old versions of FP growth and frequent item sets. You might have to update to 8.2 to get a performance bump.
Hi@JeffChowaniec,
I'm using RapidMiner 8.2.001
I tried running your process and I found that the web crawl runs for 25+ mins and I wasn't able to finish the process because I need my machine for some other tasks. I have a 32gb machine and I could see it getting taxed pretty hard at some points. Have you tried it with a data set that is a fraction of what you are trying to query? The idea is we want to make sure that even a small data set in this case will run and not take up the available memory before we dedicate a 1 hr+ run time to this.
I haven't tried to crawl less pages because once I crawled stored it in the repository. This file is to huge to upload it here. Instead I here is a repository file after the Numerical to Binominal Operator as input data for FP-growth.
Hi,
I think I've found my problem with the memory. I had to cut the FP-Growth parameter max items per itemset from 0 to 2. Now I struggle with filling the ExampleSet from the XML file "Writing Association Rules to Exampleset or file". In this example Data import wizard fullfills in Step 4 automaticly the column current value. That doesn't happens in my approach and I don't know why.
Hi,
it's me again.
I'm trying to sort out how it might be possible add the Item names in@bhupendra_patilapproachWriting-Association-Rules-to-Exampleset-or-file. The approach FP-Growth runs and I see all Columns more or less filed but if I'm using instead the new FP-Growth the Item names are not shown. Has anyone an idea how this is is possible?
If I'm using this approach on my process, than I'll see all the numerical values but no item names.
best regards
Tobias