| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 5.2 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
O Praxis é um ponto de encontro entre estudantes do ensino superior e empregadores num mercado de trabalho globalizado. Os principais fatores impulsionadores do mercado são as ofertas de estágio, submetidas pelos empregadores, e as palavras-chave das pesquisas realizadas pelos estudantes que se encontram à procura de um estágio. A preocupação principal do Praxis está relacionada com a discrepância que possa existir entre os estágios disponíveis e as cerca de 1000 pesquisas realizadas por dia. Este trabalho visa preencher eventuais lacunas existentes no mercado Praxis entre a oferta – propostas de estágio -- e a procura – palavras chave usadas nas pesquisas.
Este trabalho analisa o resultado da aplicação de diferentes técnicas de processamento automático de texto em combinação com documentos orientados a dados. Esta associação tira partido de todos os recursos existentes nos navegadores de internet mais modernos, proporcionando assim uma experiência mais agradável para o utilizador, incluindo unicamente as informações que são relevantes e que, de outra forma, passariam despercebidas para o utilizador.
Além disso, as técnicas de processamento automático de texto servem para analisar as ofertas e estágio e as palavras-chave pesquisadas, sendo que podem melhorar a correspondência entre a oferta e a procura no mercado virtual Praxis.
O trabalho Praxis Market Gap Analysis demonstra também o potencial de usar nuvens de etiquetas e técnicas de clustering em combinação com documentos orientados a dados. Esta abordagem contribui para uma melhoria das representações gráficas baseadas em dados e da qualidade da informação disponibilizada ao utilizador final.
Os principais resultados deste trabalho sugerem que os diferentes algoritmos de clustering têm resultados distintos no que respeita à organização do corpus, sendo que isto afeta a eficácia do processo de clustering. Os resultados da avaliação orientaram-nos para os algoritmos de clustering mais eficientes atendendo ao contexto e às necessidades do Praxis.
Praxis is a meeting point where higher education students meet employers in the global labor market. The main entities driving the market are the internship offers, submitted by the providers, and the keyword searches from the students looking for an internship. The core concern of the Praxis network is related to the mismatch between the internships being offered (supply) and the searches being requested (demand). Hence, this work aims to fill the gap in the Praxis market between the demand and supply. This work analyses the outcome from the application of different text mining techniques in combination with Data-Driven Documents in the Praxis network. Such combination takes advantage of the full capabilities in modern browsers and thus provides a pleasant user experience by displaying only the relevant information that was previously unseen but contained in the data. Furthermore, using text mining techniques to analyze the internship offers and the keyword searches being made in the database might improve the matching between demand and supply in the Praxis virtual market. The Praxis Market Gap Analysis portraits the potential of using tag clouds and clustering while combining Data-Driven Documents with mined data. This approach increases the number of data-driven representations and the quality of information that the final user perceives. The main findings of this study suggest that the different clustering algorithms have disparate results in regard to the organization of the input corpus which impacts the effectiveness of the clustering process. The evaluation results guided us towards the most efficient clustering algorithms for the specific context and needs of Praxis.
Praxis is a meeting point where higher education students meet employers in the global labor market. The main entities driving the market are the internship offers, submitted by the providers, and the keyword searches from the students looking for an internship. The core concern of the Praxis network is related to the mismatch between the internships being offered (supply) and the searches being requested (demand). Hence, this work aims to fill the gap in the Praxis market between the demand and supply. This work analyses the outcome from the application of different text mining techniques in combination with Data-Driven Documents in the Praxis network. Such combination takes advantage of the full capabilities in modern browsers and thus provides a pleasant user experience by displaying only the relevant information that was previously unseen but contained in the data. Furthermore, using text mining techniques to analyze the internship offers and the keyword searches being made in the database might improve the matching between demand and supply in the Praxis virtual market. The Praxis Market Gap Analysis portraits the potential of using tag clouds and clustering while combining Data-Driven Documents with mined data. This approach increases the number of data-driven representations and the quality of information that the final user perceives. The main findings of this study suggest that the different clustering algorithms have disparate results in regard to the organization of the input corpus which impacts the effectiveness of the clustering process. The evaluation results guided us towards the most efficient clustering algorithms for the specific context and needs of Praxis.
Description
Keywords
Processamento automático de texto Nuvem de etiquetas Documentos orientados a dados Clustering Text Mining Tag Clouds Data-Driven Documents