[SOLVED] Rapidminer 5.3 on Linux

spok2

Dear all,

many thanks in advance for a hint what might be wrong ...

I use a text mining process on Windows, consisting of

- Process documents from files
-- Tokenize
——过滤Stopwords
-- Stem (Porter)
-- Filter Tokens by Length
-- Transform Cases

with no problems.

Now I tried to copy the process to Ubuntu 14.04

- either by copying the *.rpm file
- or by newly creating the process on ubuntu

To my opinion both processes are identical; both tell me: "No problems found"

Starting the process on Windows I get results after about 2 minutes (Wordlist and Example Set)
Starting the process on Ubuntu I get an emty result after 0 seconds (Wordlist as well as Example set are both empty); process stops without errors.

What might go wrong?

Marco_Boeck

Hi,

my bet would be some encoding issue. Check your process XML for weird symbols.

Regards,
Marco

spok2

Marco,

thanks, but this is not the problem.

What I did (System: Ubuntu 14.04 with Oracle Java 1.8.0_31):
- completely deleted rapidminer, including .RapidMiner5 in $HOME
- downloaded rapidminer 5.3.13 from Sourceforge
- unzip below my local homedir
- set all dirs to 755, all files to 644
- copied 5 arbitrary english articles from the web and put them in a directory
- copied RapidMinerGUI from scripts to installation dir
- started rapidminer
- made update to 5.3.15
- installed extensions
- created a new process, consisting of the elements I already mentioned (xml see below)
- get no errors, as well as no results after running the process, ready within 0s

Same process runs on Windows 7 without any problem

Marco_Boeck

Hi,

有一个区别,可能是负责任的: On Linux, you're using your local Java installation, i.e. Java 8 in your case. On Windows, a JRE is shipped with RapidMiner, so you're using a Java 7 version there. You might want to try again on Linux with Java 7.

Regards,
Marco

spok2

Marco,

thanks, thatś it ...

I removed Oracle Java8, installed Oracle Java7 and all is fine.

Best regards

Wolfgang