Diagnóstico médico usando raciocínio baseado em casos e aprendizagem de máquina

Abstract

The objective of the work is to develop a methodology for diagnosing of diseases using Case-Based Reasoning (CBR). Casebase preprocessing includes data normalization, handling of missing values, and selection of significant features, contributing to an accurate casebase. The k-NN (k-Nearest Neighbors) algorithm is used to determine the proximity of past cases and identify similar patterns. Judicious choice of parameter k plays a role crucial in this process, directly influencing the sensitivity of the diagnosis. Case adaptation uses artificial neural networks and exploits the flexibility and learning capacity of neural networks to adjust the retrieved cases, taking into account consideration specific nuances of the patient in question. Dynamic adaptation improves the system's ability to deal with individual variations and complexities present in different clinical cases. The results obtained demonstrate the effectiveness of proposed approach in the diagnosis of diseases. The integration of pre-processing steps, case retrieval with k-NN and case adaptation with neural networks contributes to a robust and flexible system, capable of handling a variety of clinical scenarios