The role of named entities in web people search

Abstract

The ambiguity of person names in the Web has become a new area of interest for NLP researchers. This challenging problem has been formulated as the task of clustering Web search results (returned in response to a person name query) according to the individual they mention. In this paper we compare the coverage, reliability and independence of a number of features that are potential information sources for this clustering task, paying special attention to the role of named entities in the texts to be clustered. Although named entities are used in most approaches, our results show that, independently of the Machine Learning or Clustering algorithm used, named entity recognition and classification per se only make a small contribution to solve the problem.

Learning similarity functions for topic detection in online reputation monitoring

“…it focuses on the area where both clusters are closets to each other). A drawback of this method is that clusters may be merged due to single noisy elements being close to each other, but in practice it seems to be the best choice in problems related to ours [23,28,5]. …”

“…Following the methodology proposed in [5] for a different clustering problem, we model the problem as a binary classification task: given a pair of tweets d1, d2 , the system must decide whether the tweets belong to the same topic (true) or not (false). Each pair of tweets is represented as a set of features (for instance, term overlapping between both tweets), which are used to feed a machine learning algorithm that learns a similarity function.…”

How do humans distinguish different people with identical names on the web?

“…As named entity recognition (NER) is used in most approaches, Artiles et al investigated which document features contribute to person name disambiguation and reported that NER only makes a small contribution [4].…”

A Quantitative Evaluation of Global Word Sense Induction

“…Two evaluation metrics are employed during the unsupervised evaluation in order to estimate the quality of the clustering solutions, the V-measure [24] and the paired F-Score [25]. V-Measure assesses the quality of a clustering by measuring its homogeneity (h) and its completeness (c).…”

“…In the paired F-Score [25] evaluation, the clustering problem is transformed into a classification problem [4]. A set of instance pairs is generated from the automatically induced clusters (F(K)), which comprises pairs of the instances found in each cluster.…”