CBR with Improved Adaption and Recovery

Camilo Andrés Piñeros Rodríguez and David Ramiro Bastidas Torres, Universidad de Nariño, San Juan de Pasto-Colombia 2018

Reasoning in humans is generally based on the process of remembering and applying rules, the product of various past experiences that generate knowledge. Case-based reasoning (CBR) is a problem solving approach that uses past experience to tackle current problems.Technically, CBR is a methodology that has proven to be appropriate for applying analogy strategies in unstructured domains and where knowledge acquisition is difficult. Therefore, it is the ideal methodology for the development of support systems for medical diagnosis, through previous analysis, it can provide results that allow a better understanding of the patient and, therefore, a better diagnosis and a better treatment. The porpuse for this Project is the search for  improvement in the adaption and recovery stages with a better clasification process and more accurate probability estimation. 

Figure 1 CBR cycle

The systems based on CBR contemplate independent stages for the recovery and adaptation phase, in this work these stages are integrated into a single one, resulting in computational savings. The majority of CBR systems are built using KNN algorithm as part of the retrieve stage, and usually by the its complexity the adaption stage is avoided. This article proposes a more complex technique based on sequential classification with classifiers of different types, entering a wide research field.

Graphic Interface

Trying to improve the interaction between user and system, a more visually-friendly interface is made. Putting a more precise and a different approach towards a better classification this interface let you train, classify and analyse different cases with a variety of behaviour and forms.
A few features are:
  •  Training and test of the system classifiers.
  • Label of a new case.
  •  probability for the new case to be assigned to each class.
  • Feature Drawing. 


6th International Work-Conference on Bioinformatics and Biomedical Engineering, held at the Science Faculty of the University of Granada.


Source Code


CBR with improved adaption and recovery



 David Bastidas Camilo Piñeros
 Electronic Engineer  Electronic Engineer