Publicação
Explainable artificial intelligence on smart human mobility: a comparative study approach
| Resumo: | Explainable artificial intelligence has been used in several scientific fields to understand how and why a machine learning model makes its predictions. Its characteristics have allowed for greater transparency and outcomes in AI-powered decision-making. This building trust and confidence can be useful in human mobility research. This work provides a comparative study in terms of the explainability of artificial intelligence on smart human mobility in the context of a regression problem. Decision Tree, LIME, SHAP, and Seldon Alibi are explainable approaches to describe human mobility using a dataset generated from New York Services. Based on our results, all of these approaches present relevant indicators for our problem. |
|---|---|
| Autores principais: | Rosa, Luís |
| Outros Autores: | Silva, Fábio André Souto; Analide, Cesar |
| Assunto: | Explainable artificial intelligence Machine learning Smart cities Smart human mobility Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática Cidades e comunidades sustentáveis |
| Ano: | 2023 |
| País: | Portugal |
| Tipo de documento: | comunicação em conferência |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade do Minho |
| Idioma: | inglês |
| Origem: | RepositóriUM - Universidade do Minho |
| Resumo: | Explainable artificial intelligence has been used in several scientific fields to understand how and why a machine learning model makes its predictions. Its characteristics have allowed for greater transparency and outcomes in AI-powered decision-making. This building trust and confidence can be useful in human mobility research. This work provides a comparative study in terms of the explainability of artificial intelligence on smart human mobility in the context of a regression problem. Decision Tree, LIME, SHAP, and Seldon Alibi are explainable approaches to describe human mobility using a dataset generated from New York Services. Based on our results, all of these approaches present relevant indicators for our problem. |
|---|