I am a beginner in working with data mining, so I would really appreciate your help. I have a dataset with approximately million entries. I ran the split validation with using the j48 algorithm and still didn't get any results and the programme has been runing for 4 days already. So can anyone tell me, if I should hope for results or all of this was just in vain and there is some more efficient way to get the results.
well, decision trees are not necessarily the fastest learning schemes and it's runtime mainly depends on wisely chosen parameter settings. You could increase the level of prepruning or restrict the depth of the trees. If other learning schemes are also fine I would try different and simpler schemes first (like Naive Bayes for example).
How to load processes in XML from the forum into RapidMiner: Read this!
I really appreciate your response. I have already used the NaiveBayes and obtained results, just want to make a comparisson of the accuracy in both cases, since with J48 I get more precise results. I will take your advice and try increasing the level of prepruning and restricting the depth. Anyway thanks for your reply.