| Texto completo | |
| Autor(es): |
Nepomuceno, Pedro Ivo Siqueira
;
Braghetto, Kelly Rosa
Número total de Autores: 2
|
| Tipo de documento: | Artigo Científico |
| Fonte: | FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE; v. 177, p. 14-pg., 2026-04-01. |
| Resumo | |
Semantic heterogeneity, such as changes in grouping levels or category names, is a common challenge when analyzing long-time-lapse data collections. In previous work, we proposed a model for storing semi-structured data while tracking semantic evolution, along with querying algorithms and a prototype system (MellowDB). Initial performance tests were conducted with synthetic data. This article extends our framework by introducing a new semantic evolution operation, evaluating performance with real-world data from the Brazilian public health system, and analyzing the impact of indexing on query performance. We also propose and compare two alternative approaches for handling semantic heterogeneity: an eager one (data preprocessing) and a lazy one (query rewriting). The eager approach handles semantic evolution at insertion time, incurring processing costs upfront, whereas the lazy approach defers this cost to query time. The MellowDB source code, along with the dataset used in the performance evaluation and all associated analysis scripts, is publicly available to support transparency and reproducibility. Our results show that both strategies are viable for production use, each excelling in different scenarios. Query rewriting outperforms data preprocessing in write-heavy workloads, where insertions account for 95 % or more of operations. In all other scenarios, data preprocessing offers superior performance, highlighting it as an efficient and innovative solution for managing semantic heterogeneity in evolving datasets. (AU) | |
| Processo FAPESP: | 23/18026-8 - Centro de Ciência de Dados para Estatísticas Públicas (CCDEP) |
| Beneficiário: | Carlos Eduardo Torres Freire |
| Modalidade de apoio: | Auxílio à Pesquisa - Centros de Ciência para o Desenvolvimento |
| Processo FAPESP: | 23/00779-0 - Desenvolvimento de infraestrutura para integração e melhoria de acessibilidade de dados socioambientais e subsídio de políticas públicas |
| Beneficiário: | Kelly Rosa Braghetto |
| Modalidade de apoio: | Auxílio à Pesquisa - Regular |