I have unlabeled data that contains a hundred fields, and it is necessary to label it. One way to do it is to interpret the data and create new data that could be labeled, or find out some of those fields to label data. But doing this, we might be labeling incorrectly the data.
Are there any guidelines that we must follow that would help us in selecting a good candidates to labeled data?
You cannot train a classifier with unlabeled data. So, You need labeled examples. There are services that will label it for you, but it might be simpler for you to do it by hand.
Conference Paper Unlabeled data classification via support vector machines an...
you have to have some pre-labeled data that you will use in building your labeling (classification ) model; this what is called supervised learning. In other hand, if you don not have these pre-labeled data and you want to label the data you have, the problem now is different ; it is called unsupervised learning (clustering). The unsupervised learning is not that easy; as you should know how many clusters the data may be classified to, and there is no grantee for perfection in your results.
- Badges
- Science topic
- Similar topics
- Philosophy
- Esthetics
- Interpretation