TAILIEUCHUNG - Báo cáo khoa học: "Distributional Similarity vs. PU Learning for Entity Set Expansion"

Distributional similarity is a classic technique for entity set expansion, where the system is given a set of seed entities of a particular class, and is asked to expand the set using a corpus to obtain more entities of the same class as represented by the seeds. This paper shows that a machine learning model called positive and unlabeled learning (PU learning) can model the set expansion problem better. Based on the test results of 10 corpora, we show that a PU learning technique outperformed distributional similarity significantly. . | Distributional Similarity vs. PU Learning for Entity Set Expansion Xiao-Li Li Institute for Infocomm Research 1 Fusionopolis Way 21-01 Connexis Singapore 138632 xlli@ Bing Liu University of Illinois at Chicago 851 South Morgan Street Chicago Chicago IL 60607-7053 UsA liub@ Abstract Distributional similarity is a classic technique for entity set expansion where the system is given a set of seed entities of a particular class and is asked to expand the set using a corpus to obtain more entities of the same class as represented by the seeds. This paper shows that a machine learning model called positive and unlabeled learning PU learning can model the set expansion problem better. Based on the test results of 10 corpora we show that a PU learning technique outperformed distributional similarity significantly. 1 Introduction The entity set expansion problem is defined as follows Given a set s of seed entities of a particular class and a set D of candidate entities . extracted from a text corpus we wish to determine which of the entities in D belong to s. In other words we expand the set s based on the given seeds. This is clearly a classification problem which requires arriving at a binary decision for each entity in D belonging to s or not . However in practice the problem is often solved as a ranking problem . ranking the entities in D based on their likelihoods of belonging to s. The classic method for solving this problem is based on distributional similarity Pantel et al. 2009 Lee 1998 . The approach works by comparing the similarity of the surrounding word distributions of each candidate entity with the seed entities and then ranking the candidate entities using their similarity scores. Lei Zhang University of Illinois at Chicago 851 South Morgan Street Chicago Chicago IL 60607-7053 UsA zhang3@ See-Kiong Ng Institute for Infocomm Research 1 Fusionopolis Way 21-01 Connexis Singapore 138632 skng@ In .

• Báo cáo khoa học
• Distributional Similarity
• PU Learning
• Entity Set Expansion
• báo cáo khoa học
• báo cáo ngôn ngữ
• ngôn ngữ tự nhiên
• Measures of Distributional Similarity
• Measures of Distributional Similarity Lillian Lee
• Long Papers
• mô hình ngôn ngữ
• xử lý ngôn ngữ tự nhiên
• Verb Classification
• Syntactic and Semantic Structures
• Reducing semantic drift
• bagging and distributional similarity
• Tara McIntosh and James R
• Curran
• Scaling Distributional
• Similarity to Large Corpora
• James Gorman and James R
• Exploring Distributional
• Similarity Based Models
• Query Spelling Correction
• Directional Distributional Similarity
• Lexical Expansion
• Integrating Pattern based
• Distributional Similarity Methods
• Lexical Entailment Acquisition
• Syntax is from Mars
• Semantics from Venus! Insights
• Spectral Analysis
• Distributional Similarity Networks
• Finding Synonyms
• Automatic Word Alignment
• Distributional Similarity Models
• Clustering Neighbors
• Lillian Lee
• Word classification
• combined measures of distributional
• semantic similarity
• scientific reports
• model language
• process natural language
• Finding Word Substitutions
• a Distributional Similarity Baseline
• Immediate Context Overlap
• Information content
• Information theory
• Biomedical ontologies
• Distributional based methods
• BMC Bioinformatics
• Identifying biomedical
• Empirical risk minimization
• Approximate subgraph matching
• Distributional similarity model