A Heterogeneous Field Matching Method for Record Linkage

Minton, Steven; Nanjo, Claude; Knoblock, Craig A.; Michalowski, Martin; Michelson, Matthew

doi:10.1109/icdm.2005.7

Cited by 33 publications

(24 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The experimental results show that trainable similarity measures are capable of learning the specific notion of similarity that is appropriate for a specific domain. While this approach focuses on homogenous string-based transformation, [15] uses heterogeneous set of models to relate complex domain specific relationships between two values. One technical issue for these learningbased approach is the selection of training datasets, especially the negative instances.…”

Section: Relatedmentioning

confidence: 99%

Learning Top-k Transformation Rules

Patro

Wang

2011

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Record linkage identifies multiple records referring to the same entity even if they are not bit-wise identical. It is thus an essential technology for data integration and data cleansing. Existing record linkage approaches are mainly relying on similarity functions based on the surface forms of the records, and hence are not able to identify complex coreference records. This seriously limits the effectiveness of existing approaches. In this work, we propose an automatic method to extract top-k high quality transformation rules given a set of possibly coreferent record pairs. We propose an effective algorithm that performs careful local analyses for each record pair and generates candidate rules; the algorithm finally chooses top-k rules based on a scoring function. We have conducted extensive experiments on real datasets, and our proposed algorithm has substantial advantage over the previous algorithm in both effectiveness and efficiency.

show abstract

Section: Relatedmentioning

confidence: 99%

Learning Top-k Transformation Rules

Patro

Wang

2011

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…Traditional record matching techniques are then used over the standardized records. Minton et al [13] propose a machine learning based approach to using transformations, where labeled examples are used to learn which applications of transformations are most useful for matching a pair of records. All of these papers assume that transformations are provided as an explicit input.…”

Section: Related Workmentioning

confidence: 99%

“…While previous work [1,7,13] describes how a set of transformations can be exploited for record matching, it does not address how to identify suitable transformations in a record matching setting. For a real-world record matching task, hundreds of string transformations could be relevant and it is a challenging task for a programmer to compile the set of relevant transformations.…”

Section: Introductionmentioning

confidence: 99%

“…Prior work [1,7,13] has recognized the importance of such representational variations for record matching and has proposed several techniques for incorporating a priori knowledge of such variations into record matching process. They have also demonstrated that exploiting variations using these techniques can significantly improve the quality of record matching.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Learning string transformations from examples

2009

View full text Add to dashboard Cite

ABSTRACT"Robert" and "Bob" refer to the same first name but are textually far apart. Traditional string similarity functions do not allow a flexible way to account for such synonyms, abbreviations and aliases. Recently, string transformations have been proposed as a mechanism to make matching robust to such variations. However, in many domains, identifying an appropriate set of transformations is challenging as the space of possible transformations is large. In this paper, we investigate the problem of leveraging examples of matching strings to learn string transformations. We formulate an optimization problem where we are required to learn a concise set of transformations that explain most of the differences. We propose a greedy approximation algorithm for this NP-hard problem. Our experiments over real-life data illustrate the benefits of our approach.

show abstract

“…The use of supervised (training-based) approaches or learners aims at automating the process of entity matching to reduce the required manual effort. Training-based approaches, e.g., Naïve Bayes [49], logistic regression [46], Support Vector Machine (SVM) [11,43,49] or decision trees [63,29,49,53,54,56] have so far been used for some subtasks, e.g., determining suitable parameterizations for matchers or adjusting combination functions parameters (weights for matchers, offsets). However, training-based approaches require suitable training data and providing such data typically involves manual effort.…”

Section: Combination Of Matchersmentioning

confidence: 99%

Frameworks for entity matching: A comparison

Köpcke

Rahm

2010

Data & Knowledge Engineering

346

204

View full text Add to dashboard Cite

a b s t r a c tEntity matching is a crucial and difficult task for data integration. Entity matching frameworks provide several methods and their combination to effectively solve different match tasks. In this paper, we comparatively analyze 11 proposed frameworks for entity matching. Our study considers both frameworks which do or do not utilize training data to semiautomatically find an entity matching strategy to solve a given match task. Moreover, we consider support for blocking and the combination of different match algorithms. We further study how the different frameworks have been evaluated. The study aims at exploring the current state of the art in research prototypes of entity matching frameworks and their evaluations. The proposed criteria should be helpful to identify promising framework approaches and enable categorizing and comparatively assessing additional entity matching frameworks and their evaluations.

show abstract

A Heterogeneous Field Matching Method for Record Linkage

Cited by 33 publications

References 9 publications

Learning Top-k Transformation Rules

Learning Top-k Transformation Rules

Learning string transformations from examples

Frameworks for entity matching: A comparison

Contact Info

Product

Resources

About