With the development of telecom business, customer churn prediction becomes more and more important. An outstanding issue in customer churn prediction is high dimensional problem....
We introduce a novel machine learning framework based on recursive autoencoders for sentence-level prediction of sentiment label distributions. Our method learns vector space repr...
Richard Socher, Jeffrey Pennington, Eric H. Huang,...
The intuition behind ensembles is that different prediciton models compensate each other’s errors if one combines them in an appropriate way. In case of large ensembles a lot of...
Krisztian Buza, Alexandros Nanopoulos, Lars Schmid...
The goal of sentiment prediction is to automatically identify whether a given piece of text expresses positive or negative opinion towards a topic of interest. One can pose sentim...
In this paper, we investigate the problem of automatically predicting segment boundaries in spoken multiparty dialogue. We extend prior work in two ways. We first apply approaches...