Entity-Based Retrieval in Shared Semi-Structured Information Spaces

Proceedings of CIKIM. 1996
Entity-Based Retrieval in Shared Semi-Structured Information Spaces

Semi-structured information sharing systems are gaining in popularity because they allow users to easily create shared collections of textual documents, organized by a common set of fields. Unfortunately, in a large organization this freedom can result in an unwieldy space of shared information that is difficult to retrieve.

Standard tools like full-text search do not alleviate the problem, in part because they do not make any use of the structure within each document collection. In this paper, we describe an approach that goes beyond full-text search by taking advantage of both the structure of the document collections and a knowledge of what information types are important within the organization sharing the information.

We present an implemented indexing/browsing system called Notes Explorer that allows users to browse for entities (companies, people, etc.) across a large semi-structured information space. Notes Explorer incorporates three key components:

(1) automatic classification of document fields to recognize common entity and document collection types;

(2) entity-based browsing over multiple document collections, with type-dependent normalization;


(3) content-based filtering of browse results.

Another publication from the same category: Machine Learning and Data Science

IEEE Computing Conference 2018, London, UK

Regularization of the Kernel Matrix via Covariance Matrix Shrinkage Estimation

The kernel trick concept, formulated as an inner product in a feature space, facilitates powerful extensions to many well-known algorithms. While the kernel matrix involves inner products in the feature space, the sample covariance matrix of the data requires outer products. Therefore, their spectral properties are tightly connected. This allows us to examine the kernel matrix through the sample covariance matrix in the feature space and vice versa. The use of kernels often involves a large number of features, compared to the number of observations. In this scenario, the sample covariance matrix is not well-conditioned nor is it necessarily invertible, mandating a solution to the problem of estimating high-dimensional covariance matrices under small sample size conditions. We tackle this problem through the use of a shrinkage estimator that offers a compromise between the sample covariance matrix and a well-conditioned matrix (also known as the "target") with the aim of minimizing the mean-squared error (MSE). We propose a distribution-free kernel matrix regularization approach that is tuned directly from the kernel matrix, avoiding the need to address the feature space explicitly. Numerical simulations demonstrate that the proposed regularization is effective in classification tasks.