Titulo Estágio
Relationship Discovery on the Web
Áreas de especialidade
Sistemas de Informação
Local do Estágio
DEI-FCTUC
Enquadramento
The vision statement “Locating Your Next Strategic Opportunity”, published by Harvard Business Review in March 2011, clearly shows some of the innovations that can be explored by using service networks. For example, service networks can be used to identify niche markets, which are fertile to explore new ideas. Unfortunately, no attempts have been made to reconstruct such large-scale networks.
This project aims to develop an approach to discover relationships between service providers on the web. This can be generalized as: “given the names of two service providers, discover the possible relationships that exist between them using the web as a knowledge base”.
Our idea to discover relationships is to use web search engines (e.g., Google) and information retrieval methods to find evidence of relations between services.
Objetivo
We will rely on co-occurrence analysis to discover relationships. The method uses web hits for measuring the association between terms using the overlap (Simpson) coefficient, which is calculated based on the number of web hits for each individual term and their conjunction (i.e., AND query of the two terms). The co-occurrence of two service names generally indicated the existence of a relationship between services.
For example, we will discover a relationship between a service A (e.g.,”SalesNexus", a CRM cloud service provider) and a service B (e.g., “Maximizer", also a CRM cloud service) by submitting a boolean query combining the two service names (e.g., “SalesNexus" and “Maximizer"). A term similar to “CRM OR Enterprise System" will be added to the query for disambiguation and a term similar to “provider OR competitor OR customer” will be added to restrict the search to web pages that mention possible relationships.
The method will be extended to use natural language processing (NLP), text analysis, and computational linguistics to identify and extract subjective information from web pages discussing possible relationships between services.
Plano de Trabalhos - Semestre 1
(a) Study the main technologies to be used by the project (e.g., NLP, Google, JSON, Web API, Semantic Web) (Setembro a Dezembro de 2014).
(b) Develop the first prototype of a discovery engine to discover service relationships (Minimum Viable Product) (Novembro de 2014 a Fevereiro de 2015).
Plano de Trabalhos - Semestre 2
(c) Testes the discovery engine with real data (Março de 2015).
(d) Develop the final system prototype of a discovery engine for service networks (Março de 2014 a Jun de 2015).
(e) Documentation (running task from begin to end) (Setembro 2014 a Jun de 2015).
(f) Report writing and defense (Abril a Jun de 2015).
Condições
Este trabalho será realizado no DEI/Universidade de Coimbra.
Mestrado não remunerado.
Observações
If the results of the project are complete, sound, and innovative, it will be possible to write a final research paper to be published in a scientific outlet.
Orientador
Jorge Cardoso
jcardoso@dei.uc.pt 📩