The index should start from 1 for LIBSVM format, as defined in the README of LIBSVM (https://github.com/cjlin1/libsvm/blob/master/README#L64). The only exception is the precomputed kernel, which MLlib doesn't support. -Xiangrui
On Wed, May 6, 2015 at 1:42 AM, doyere <doy...@doyere.cn> wrote: > Hi all, > > After do some tests,finally I solve it.I wrote here for other people who > met this question. here's a example of data format error I faced > > 0 0:0 1:0 2:1 > 1 1:1 3:2 > > the data for 0:0 and 1:0/1:1 is the reason for > ArrayIndexOutOfBoundsException.If someone who faced the same question just > delete them from u past data or update it. Since it was worked in > libsvm-tools,so I guess in spark MLlib it just implements a bit different. > > 原始邮件 > *发件人:* doyere<doy...@doyere.cn> > *收件人:* user<user@spark.apache.org> > *发送时间:* 2015年5月6日(周三) 08:59 > *主题:* MLlib libsvm isssues with data > > hi all: > > I’ve met a issues with MLlib.I used posted to the community seems put the > wrong place:( .Then I put in stackoverflowf.for a good format details plz > see > http://stackoverflow.com/questions/30048344/spark-mllib-libsvm-isssues-with-data.hope > someone could help [image: 😢] > > I guess it’s due to my data.but I’ve test it in libsvm-tools it worked > well,and I’ve used the libsvm data python data format test tool and it’s > ok.Just don’t know why it errors with java.lang. > ArrayIndexOutOfBoundsException: -1 :( > > And this is my first time using the mail list ask for help.If I did > something wrong or I described not clearly plz tell me. > > > doye >