A carregar...
Publicação
A criação de um modelo de Natural Language Processing para extração de habilidades técnicas na área de Ciência de Dados
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.23 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
O trabalho surgiu da necessidade do homem de suprir suas necessidades básicas. Na
antiguidade, época de Gregos e Romanos, era ensinado a prole como cuidar da terra e esse
conhecimento era passado por gerações. Com a chegada da primeira revolução industrial e
posteriormente com a popularização dos computadores, o trabalho se tornou o que
conhecemos hoje. A forma de contratação mudou ao longo dos anos e com a chegada da
tecnologia mais do que nunca, o mercado de trabalho está em constante mudança e como
consequência disto as habilidades necessárias para estes trabalhos também seguem esta
mudança. Grande parte das vagas de trabalho hoje estão anunciadas em sites de buscas de
emprego como Indeed, Glassdoor e Linkedin. Um grande desafio atual, é prever as mudanças
e tendências do mercado de trabalho em termos de habilidades, e a analise textual pode gerar
uma vantagem competitiva neste sentido.
A proposta deste trabalho é analisar através de técnicas de Natural Language Processing
(NLP) diferentes oportunidades de emprego da área de Data Science a fim de obter um
modelo que possa ser utilizado para extrair as habilidades requisitadas para esta área.
Para alcançar este objetivo primeiro é feita uma revisão dos conceitos de Aprendizado de
Máquina, Natural Language Processing, Transfer Learning e das técnicas de preparação de
dados, e em seguida será apresentada a metodologia utilizada. Depois, são destacadas as
técnicas que funcionam melhor para extração de habilidades, a escolha e criação do modelo, e
por fim a apresentação de resultados.
The work originated from man's need to meet his basic needs. In ancient times, the times of the Greeks and Romans, offspring were taught how to take care of the land and this knowledge was passed on for generations. With the arrival of the first industrial revolution and later with the popularization of computers, work became what we know today. The way of hiring has changed over the years and with the arrival of technology more than ever, the job market is constantly changing, and consequently, the skills needed for these jobs also follow this change. A large part of job vacancies today is advertised on job search sites such as Indeed, Neuvoo, and Linkedin. A big challenge today is to predict the changes and trends in the job market in terms of skills and textual analysis can generate a competitive advantage in this sense. The purpose of this work is to analyze through Natural Language Processing (NLP) techniques different job opportunities in the Data Science area to obtain a model that can be used to extract the required skills for this area. To achieve this goal first a review of the concepts of Machine Learning, Natural Language Processing, Transfer Learning, and data preprocessing, then the methodology used is presented. Next, the techniques that work best for skill extraction is highlighted, the choice and creation of the model, and finally the presentation of results.
The work originated from man's need to meet his basic needs. In ancient times, the times of the Greeks and Romans, offspring were taught how to take care of the land and this knowledge was passed on for generations. With the arrival of the first industrial revolution and later with the popularization of computers, work became what we know today. The way of hiring has changed over the years and with the arrival of technology more than ever, the job market is constantly changing, and consequently, the skills needed for these jobs also follow this change. A large part of job vacancies today is advertised on job search sites such as Indeed, Neuvoo, and Linkedin. A big challenge today is to predict the changes and trends in the job market in terms of skills and textual analysis can generate a competitive advantage in this sense. The purpose of this work is to analyze through Natural Language Processing (NLP) techniques different job opportunities in the Data Science area to obtain a model that can be used to extract the required skills for this area. To achieve this goal first a review of the concepts of Machine Learning, Natural Language Processing, Transfer Learning, and data preprocessing, then the methodology used is presented. Next, the techniques that work best for skill extraction is highlighted, the choice and creation of the model, and finally the presentation of results.
Descrição
Dissertation presented as the partial requirement for obtaining a Master's degree in Data Science and Advanced Analytics, specialization in Data Science
Palavras-chave
Aprendizagem de Máquina Aprendizagem Profunda Habilidades Habilidades Técnicas Inteligência Artificial Machine Learning Skills Deep Learning Technical Skills Artificial Intelligence