Resumo

CASAS MIRANDA, Roberto Jose María et al. Image classification with neural networks improves consumer behavior in college students. Revista InveCom [online]. 2026, vol.6, n.2, e602093.  Epub 30-Set-2025. ISSN 2739-0063.  https://doi.org/10.5281/zenodo.17027126.

The study aimed to determine how image classification using neural networks (NN) improves consumer behavior among university students. An applied research study was proposed, with a pre-experimental design and explanatory level. Data was collected and processed using non-participatory techniques, analyzing the digital interaction of university students. The data was cleaned and organized for analysis, in which machine learning techniques were applied. The use of TensorFlow and TensorFlow Datasets simplified the preprocessing and training of the model, ensuring efficient flow and optimized performance through image normalization and data batch configuration. The model used detailed visualizations with clear indicators to analyze its performance, highlighting successes and errors, which facilitated its interpretation and adjustment. The pre-trained VGG16 model, adjusted in 15 epochs, achieved a validation accuracy of 97.4%, demonstrating its high effectiveness on unseen data. The results highlight the transformative impact of CNNs on product classification, improving the user experience and optimizing consumer satisfaction and perception on digital fashion platforms. The implementation of the RN-based image classifier significantly improved the user experience, shifting from a negative perception to 90% positive ratings. The Wilcoxon test confirmed that this change reflects a real improvement in the behavior and decisions of college students as consumers.

Palavras-chave : neural networks; image classifier; machine learning.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )