Implementación de un olfatómetro electrónico portátil para la detección no invasiva de Ehrlichia Canis y evaluación de los datos a través de métodos de aprendizaje automático.

Abstract

The main objective of this internship was to implement a non-invasive and novel method for the detection of canine Ehrlichiosis. This method consisted of analyzing by means of an olfactometer composed of commercial MQ gas sensors, the volatile organic compounds (VOCs) generated by the breath samples of canines infected with the disease and a control group (healthy canines), to make the comparison. Additionally, with the use of pattern recognition and machine learning methods, we sought to obtain a discrimination and classification of infected patients in comparison with the group of controls. The search for a low-cost alternative method accessible to all is of great interest since traditional tests such as serology and PCR, although effective, are also expensive and their results require time; early detection is crucial to establish an effective treatment and prevent a fatal disease in canines.