A carregar...
Publicação
ChatGPT as a Data Scientist: Can AI Handle Clustering Better Than a Human?
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 4.66 MB | Adobe PDF |
Autores
Resumo(s)
Artificial intelligence (AI) has rapidly advanced in recent years, becoming a significant trend in various fields, including data science. This thesis explores how ChatGPT, an AI language model, performs as a data scientist compared to a human counterpart. Using both quantitative and qualitative methods, the study evaluates ChatGPT's ability to handle essential data science tasks like code generation, code completion, and clustering analysis. Insights from a survey completed by data scientists in Portugal shed light on their experiences with AI tools and their views on AI's role in their work. The research also involved a hands-on comparison between ChatGPT and a human data scientist handling a clustering problem. The results show that while ChatGPT is efficient and capable in managing routine data science tasks, it struggles with more complex analyses and often requires human oversight. Ethical concerns, such as biases in AI outputs, were also addressed. The findings suggest that AI and human expertise complement each other in the field of data science. Although AI tools like ChatGPT can boost productivity and efficiency, they cannot entirely replace the nuanced decision-making and creative problem-solving skills of human data scientists. This research emphasizes the importance of continued development to overcome these limitations and better integrate AI into data science workflows.
Descrição
Dissertation presented as the partial requirement for obtaining a Master's degree in Information Management, specialization in Knowledge Management and Business Intelligence
Palavras-chave
ChatGPT Data Science Artificial Intelligence Clustering Python