Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I am really interested in learning techniques which can help me deal with the problem briefly described below.

I have a two-class {$0,1$} classification problem.

Most of the object attributes are incremental in time: when new object appears we know very little about it; as time passes by its "incremental" dynamic attributes are being updated.
Most of these "dynamic" features are incremental - for example such as

  • number of visitors of the web page
  • number of times the product was purchased etc.

Objects arrive in discrete time. Object's label is quite expensive - in the sense that it is rather difficult for expert to classify newly appeared object.

False Positive $(\widehat{y}=1, y=0)$ is VERY BAD, while False Negative $(\widehat{y}=0, y=1)$ is not so much.

Main goal is to find some classification model, which achieves good balance between

  1. model's high True Positive Rate with really low False Positive Rate and

  2. time it takes to accumulate object's dynamic attributes for its confident classification by the model.

It is crucial to classify new object $(X,Y)$ as soon as possible if its class label is $1(Y=1)$.

share|improve this question

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.