Label Propagation through Linear Neighborhoods

• Published in: IEEE transactions on knowledge and data engineering Vol. 20; no. 1; pp. 55 - 67
• Main Authors: Wang, Fei, Zhang, Changshui
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.01.2008; IEEE Computer Society; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Applied sciences; Artificial intelligence; Classification; Computer science; control theory; systems; Data mining; Data processing. List processing. Character string processing; Exact sciences and technology; Gaussian processes; Graph labeling; Labels; Learning; Machine learning; Machine learning algorithms; Memory organisation. Data processing; Mining methods and algorithms; Predictive models; Semifabricated products; Semisupervised learning; Software; Speech and sound recognition and synthesis. Linguistics; Studies; Supervised learning; Support vector machine classification; Support vector machines; Tasks; Text categorization; Texts; Unsupervised learning; Data mining; Graph labeling; Mining methods and algorithms; Machine learning; Data analysis; semisupervised learning; label propagation; Information retrieval; Information extraction; Text; Linear model; graph; Structured data mining; Natural language; Data structure; linear neighborhoods; Learning algorithm; Artificial intelligence; Content analysis
• ISSN: 1041-4347; 1558-2191
• DOI: 10.1109/TKDE.2007.190672

In many practical data mining applications such as text classification, unlabeled training examples are readily available, but labeled ones are fairly expensive to obtain. Therefore, semi supervised learning algorithms have aroused considerable interests from the data mining and machine learning fields. In recent years, graph-based semi supervised learning has been becoming one of the most active research areas in the semi supervised learning community. In this paper, a novel graph-based semi supervised learning approach is proposed based on a linear neighborhood model, which assumes that each data point can be linearly reconstructed from its neighborhood. Our algorithm, named linear neighborhood propagation (LNP), can propagate the labels from the labeled points to the whole data set using these linear neighborhoods with sufficient smoothness. A theoretical analysis of the properties of LNP is presented in this paper. Furthermore, we also derive an easy way to extend LNP to out-of-sample data. Promising experimental results are presented for synthetic data, digit, and text classification tasks.