ZZ question: Solve the weka jvm outofmemory problem [solved]

Sender: finallyliuyu (lead donkey), email area: NLP
Question: solutions to weka jvm outofmemory Problems
Mail station: Shui mu Community (Sun Feb 20 10:36:12 2011), Station

We are using WEKA for text classification experiments (mainly focusing on measuring Feature Word selection)Algorithm).
Corpus: reteurs21578
Featurewords 100
Training set: 6950
Test Set: 2676
(In sparse data storage mode) ARFF file size: traincorpus. ARFF (538 K), testcorpus. ARFF (208 K) classifier uses SMO
The error "run outof memory" is returned.
The maxheap = 2048 m in runweka. INI is still invalid by referring to the method of one student ..
Do you have any other solutions?
I can see some methods to change the JVM size on the network, but I am not familiar with Java and do not know how to modify it. For detailed steps.
--
One NLP amateur

※Source: · shuimu community http://newsmth.net/· ~from: 159.226.20. *]

[Full text] [reply to article] [Author: finallyliuyu] [reply to author] [go to Forum] [go Back to Top] [shareSohu Weibo]

2 Sender: finallyliuyu (lead donkey), email area: NLP
Question: Re: solutions to weka jvm outofmemory Problems
Mailing site: shuimu community (Sun Feb 20 10:47:05 2011), within the site

I have uploaded my own data. I am very grateful for the help of my computer with WEKA kids shoes. In addition, I am skeptical that my WEKA installation is faulty. There are two types of WEKA on WEKA Official Website: one is built-in JVM and the other is without JVM. I have installed Java SDK on my computer before, so I chose not to bring JVM. In this way, the installed Java cannot be started on the shortcut menu.

[Mentioned in the finallyliuyu masterpiece :]
: We are using WEKA for text classification experiments (mainly focusing on measuring the advantages and disadvantages of Feature Word selection algorithms ).
: Corpus: reteurs21578
: Featurewords 100
:...................

--
One NLP amateur

※Source: · shuimu community http://newsmth.net/· ~from: 159.226.20. *]

Attachment: traincorpus. ARFF (538kb)

Attachment: testcorpus. ARFF (207kb)

[Full text] [reply to article] [Author: xwild] [reply to author] [go to Forum] [Back to Top] [shareSohu Weibo]

3 Sender: xwild (xwild), email area: NLP
Question: Re: solutions to weka jvm outofmemory Problems
Mailing site: shuimu community (Sun Feb 20 10:53:01 2011), within the site

You can try the WEKA command line and set the maximum memory.
[Mentioned in the finallyliuyu masterpiece :]
: We are using WEKA for text classification experiments (mainly focusing on measuring the advantages and disadvantages of Feature Word selection algorithms ).
: Corpus: reteurs21578
: Featurewords 100
:...................

--

※Source: · shuimu community http://newsmth.net/· ~from: 211.87.234. *]

[Full text] [reply to article] [Author: finallyliuyu] [reply to author] [go to Forum] [go Back to Top] [shareSohu Weibo]

4 Sender: finallyliuyu (lead donkey), email area: NLP
Question: Re: solutions to weka jvm outofmemory Problems
Mailing site: shuimu community (Sun Feb 20 11:28:23 2011), within the site

I am calling the WEKA Command Line Method in wekadoc, but it does not respond and does not go out of the classification process.
[Mentioned in the masterpiece xwild (xwild :]
: You can try the WEKA command line and set the maximum memory.

--
One NLP amateur

※Source: · shuimu community http://newsmth.net/· ~from: 159.226.20. *]

The image of this topic is as follows: wekasimplecli .jpg (11kb)

[Full text] [reply to article] [Author: xphan] [reply to author] [go to Forum] [Back to Top] [shareSohu Weibo]

5. Sender: xphan (dblue), email area: NLP
Question: Re: solutions to weka jvm outofmemory Problems
Mailing site: shuimu community (Sun Feb 20 16:43:30 2011), within the site

Java can use-XMS and-xmx in the command line to set the minimum and maximum memory, for example
Java-xms256m-xmx5 g...
However, when running in windows, the system also has restrictions on memory.

[Mentioned in the finallyliuyu masterpiece :]
: We are using WEKA for text classification experiments (mainly focusing on measuring the advantages and disadvantages of Feature Word selection algorithms ).
: Corpus: reteurs21578
: Featurewords 100
:...................

--

※Source: · shuimu community http://newsmth.net/· ~from: 125.39.171. *]

[Full text] [reply to article] [Author: fancyerii] [reply to author] [go to Forum] [go Back to Top] [shareSohu Weibo]

6. Sender: fancyerii (Meow's microphone), email area: NLP
Question: Re: solutions to weka jvm outofmemory Problems
Mailing site: shuimu community (Sun Feb 20 18:14:10 2011), within the site

You can debug the problem yourself.CodeOr the maillist sent to them is better.
[Mentioned in the finallyliuyu masterpiece :]
: We are using WEKA for text classification experiments (mainly focusing on measuring the advantages and disadvantages of Feature Word selection algorithms ).
: Corpus: reteurs21578
: Featurewords 100
:...................

--

※Source: · shuimu community http://newsmth.net/· ~from: 162.105.200. *]

[Full text] [reply to article] [Author: finallyliuyu] [reply to author] [go to Forum] [go Back to Top] [shareSohu Weibo]

7. Sender: finallyliuyu (lead donkey), email area: NLP
Question: Re: solutions to weka jvm outofmemory Problems
Mailing station: shuimu community (Mon Feb 21 21:14:03 2011), within the station

Thanks to xwild, this problem has been solved. The solution is as follows:
In the WEKA simplecli window, enter
Java-xmx 1024 m
Java WEKA. classifiers. Lazy. ibk-K 5-t data/traincorpus. ARFF-t data/testcorpus. ARFF-I
Here, KNN is used as an example to illustrate how to use the console to solve the WEKA memory limit problem.
References Chapter 1 Section 1 a command line primer of wekamanual provided by WEKA
Morgan. Kaufmann data mining chapter 1 Section 1 "the command line interface"

[Mentioned in the masterpiece xwild (xwild :]
: You can try the WEKA command line and set the maximum memory.

--
One NLP amateur

※source: · shuimu community http://newsmth.net/?from: 159.226.20. *]