Various semi-supervised learning methods have been proposed recently to solve the long-standing shortage problem of manually labeled data in sentiment classification. However, most existing studies assume the balance between negative and positive samples in both the labeled and unlabeled data, which may not be true in reality. In this paper, we investigate a more common case of semi-supervised ...

We consider the task of learning a classifier from the feature space X to the set of classes Y = {0, 1}, when the features can be partitioned into class-conditionally independent feature sets X 1 and X 2. We show the surprising fact that the class-conditional independence can be used to represent the original learning task in terms of 1) learning a classifier from X 2 to X 1 and 2) learning the...

The paper argues that a part of the current statistical discussion is not based on the standard firm foundations of the field. Among the examples we consider are prediction into the future, semi-supervised classification, and causality inference based on observational data.

Co-training is a well-known semi-supervised learning technique that applies two basic learners to train the data source, which uses the most confident unlabeled data to augment labeled data in the learning process. In the paper, we use the diversity of class probability estimation (DCPE) between two learners and propose the DCPE co-training approach. The key idea is to use DCPE to predict label...

Graph-based approaches have been successful in unsupervised and semi-supervised learning. In this paper, we focus on the real-world applications where the same instance can be represented by multiple heterogeneous features. The key point of utilizing the graph-based knowledge to deal with this kind of data is to reasonably integrate the different representations and obtain the most consistent m...

Semi-supervised learning (SSL) is focused on learning from labeled and unlabeled data by incorporating structural and statistical information of the available unlabeled data. The amount of data is dramatically increasing, but few of them are fully labeled, due to cost and time constraints. This is even more challenging for non-vectorial, proximity data, given by pairwise proximity values. Only ...

To select unlabeled example effectively and reduce classification error, confidence estimation for graphbased semi-supervised learning (CEGSL) is proposed. This algorithm combines graph-based semi-supervised learning with collaboration-training. It makes use of structure information of sample to calculate the classification probability of unlabeled example explicitly. With multi-classifiers, th...

In opinion mining of product reviews, one often wants to produce a summary of opinions based on product features/attributes. However, for the same feature, people can express it with different words and phrases. To produce a meaningful summary, these words and phrases, which are domain synonyms, need to be grouped under the same feature group. This paper proposes a constrained semisupervised le...

Several semi-supervised learning methods have been proposed to leverage unlabeled data, but imbalanced class distributions in the data set can hurt the performance of most algorithms. In this paper, we adapt the new approach of contrast classifiers for semi-supervised learning. This enables us to exploit large amounts of unlabeled data with a skewed distribution. In experiments on a speech act ...