ESTUDIO COMPARATIVO DE LOS ALGORITMOS DE APRENDIZAJE AUTOMÁTICO PARA EL RECONOCIMIENTO DE IMÁGENES

Abstract

The article presents the objectives of comparing the effectiveness and efficiency of various machine learning algorithms in the image recognition task, to facilitate the selection of the most optimal ones according to specific applications. Among the main theoretical contributions is the conceptualization of image recognition and the operation of algorithms such as Convolutional Neural Networks (CNN), Support Vector Machines (SVM) and Decision Trees. The methodology consisted of using a standardized data set with images labeled in several categories, divided into training, validation and test groups. The models were trained and evaluation metrics were monitored during validation and then the results in the tests were compared using statistical analysis. The results showed significant differences between the algorithms, highlighting the superior performance of the CNNs that reached 94% accuracy in the task. It is concluded that CNNs are more necessary for image recognition in relation to other algorithms, while RNN and K-Means exhibit lower performance, constituting valuable findings for future research and applications in this field.