Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Automatic Construction of Multi-faceted User Profiles using Text Clustering and its Application to Expert Recommendation and Filtering Problems

Luis M. de Campos, Juan M. Fernández-Luna, Juan F. Huete, Luis Redondo-Expósito

TL;DR

This work tackles expert finding and document filtering by automatic construction of multi-faceted user profiles through text clustering. It presents local and global clustering schemes to uncover latent topics from documents, generating subprofiles that are fused into an overall ranking for experts. Evaluated on a parliamentary corpus, the approach improves over baselines and reveals when global versus local clustering is advantageous, as well as how the number of clusters and algorithm choice impact performance. The findings demonstrate that clustering-based subprofiles can more accurately represent user interests and enhance information access tasks with practical implications for large-scale, topic-rich domains. Overall, the paper provides a coherent, device-agnostic framework for building and using multi-faceted profiles in expert recommendation and document filtering.

Abstract

In the information age we are living in today, not only are we interested in accessing multimedia objects such as documents, videos, etc. but also in searching for professional experts, people or celebrities, possibly for professional needs or just for fun. Information access systems need to be able to extract and exploit various sources of information (usually in text format) about such individuals, and to represent them in a suitable way usually in the form of a profile. In this article, we tackle the problems of profile-based expert recommendation and document filtering from a machine learning perspective by clustering expert textual sources to build profiles and capture the different hidden topics in which the experts are interested. The experts will then be represented by means of multi-faceted profiles. Our experiments show that this is a valid technique to improve the performance of expert finding and document filtering.

Automatic Construction of Multi-faceted User Profiles using Text Clustering and its Application to Expert Recommendation and Filtering Problems

TL;DR

This work tackles expert finding and document filtering by automatic construction of multi-faceted user profiles through text clustering. It presents local and global clustering schemes to uncover latent topics from documents, generating subprofiles that are fused into an overall ranking for experts. Evaluated on a parliamentary corpus, the approach improves over baselines and reveals when global versus local clustering is advantageous, as well as how the number of clusters and algorithm choice impact performance. The findings demonstrate that clustering-based subprofiles can more accurately represent user interests and enhance information access tasks with practical implications for large-scale, topic-rich domains. Overall, the paper provides a coherent, device-agnostic framework for building and using multi-faceted profiles in expert recommendation and document filtering.

Abstract

In the information age we are living in today, not only are we interested in accessing multimedia objects such as documents, videos, etc. but also in searching for professional experts, people or celebrities, possibly for professional needs or just for fun. Information access systems need to be able to extract and exploit various sources of information (usually in text format) about such individuals, and to represent them in a suitable way usually in the form of a profile. In this article, we tackle the problems of profile-based expert recommendation and document filtering from a machine learning perspective by clustering expert textual sources to build profiles and capture the different hidden topics in which the experts are interested. The experts will then be represented by means of multi-faceted profiles. Our experiments show that this is a valid technique to improve the performance of expert finding and document filtering.
Paper Structure (22 sections, 1 equation, 7 figures, 5 tables)

This paper contains 22 sections, 1 equation, 7 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Preliminaries
  3. User Profiling
  4. Text Clustering
  5. Building multi-faceted profiles by clustering documents
  6. Evaluation
  7. Test Collection
  8. Overview of the Recommender and Filtering System
  9. Clustering Algorithms
  10. Selecting the Number of Clusters
  11. Experimental context
  12. Results
  13. RQ1: Is text clustering a suitable technique to automatically extract topics in a parliamentary context?
  14. Individual Qualitative Analysis
  15. Overall Qualitative Analysis
  16. ...and 7 more sections

Figures (7)

  • Figure 1: Steps in the text clustering process
  • Figure 2: Example of local and global clustering
  • Figure 3: Word Cloud representation for the different profiles: left graph shows Monolithic (all the interventions forms a unique profile) and right graph shows a Committee-based profile obtained using the data from the "Gender Equality and Social Welfare Committee.
  • Figure 4: Word Cloud representation for two different learned profiles using Global K-Means as clustering approach.
  • Figure 5: Distribution of committees in clusters given by the Global K-Means algorithm
  • ...and 2 more figures