Ivan Gaytán Aguilar
Ivan Gaytán Aguilar has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Everardo Efrén Granda Gutiérrez
Universidad Autónoma del Estado de México
2 shared publications - Roberto Alejo
Instituto Tecnológico de Toluca
2 shared publications - Alejandro Aguilar Contreras1 shared publication
- Eréndira Rendón Lara
Tecnológico Nacional de México, Instituto Tecnológico de Toluca, Avenida Tecnológico s/n, Colonia Agrícola Bellavista, Metepec 52149, Estado de México, Mexico
1 shared publication - Federico Del Razo López
Division of Graduate Studies and Research, National Technological of Mexico, Campus Toluca, Metepec 52149, Mexico
1 shared publication - Grisel Miranda Piña
División de Ingeniería en Sistemas Computacionales, Tecnológico de Estudios Superiores de Jocotitlán, Carretera Toluca-Atlacomulco km 44.8, Ejido de San Juan y San Agustin, Jocotitlán 10587, Mexico
1 shared publication
Latest Publications
- Open Access
A Data Science Framework for Municipal Solid Waste Systems Based on Behavioral Segmentation
Municipal solid waste management (MSWM) systems in Latin America are constrained by limited access to high-resolution operational data, compelling local authorities to depend on aggregated national statistics that are inadequate for behaviorally informed intervention design. This limitation is particularly evident in the State of Mexico, which generates about 16,187 tons of waste every day but only recycles only 11%. In this context, this study introduces a diagnostic data science framework to identify behaviorally grounded citizen segments and their defining attributes, supporting evidence-based decision-making in MSWM. Primary survey data from 560 households across three municipalities were used, and a three-stage analytical pipeline was implemented to account for contextual heterogeneity. First, k-means clustering was applied to identify behavioral segments. Second, random forest classifiers were used to validate cluster coherence and quantify feature importance. Third, the Apriori algorithm was used to extract association rules that capture recurrent material-mixing behaviors. The results revealed municipality-specific segmentation structures (Tequixquiac: K = 6; Tlalpujahua: K = 3; Xalatlaco: K = 2), with material-specific disposal behaviors emerging as stronger segmentation drivers. Random forest classifiers validated cluster coherence with 100% accuracy, confirming that segments represent behaviorally distinct archetypes. The proposed framework converts raw behavioral data into actionable municipal visions. This approach focuses on finding diagnostic patterns instead of making predictions by utilizing machine-learning-driven MSWM research.
Open AccessA Data Science Framework for Municipal Solid Waste Systems Based on Behavioral Segmentation
Municipal solid waste management (MSWM) systems in Latin America are constrained by limited access to high-resolution operational data, compelling local authorities to depend on aggregated national statistics that are inadequate for behaviorally informed intervention design. This limitation is particularly evident in the State of Mexico, which generates about 16,187 tons of waste every day but only recycles only 11%. In this context, this study introduces a diagnostic data science framework to identify behaviorally grounded citizen segments and their defining attributes, supporting evidence-based decision-making in MSWM. Primary survey data from 560 households across three municipalities were used, and a three-stage analytical pipeline was implemented to account for contextual heterogeneity. First, k-means clustering was applied to identify behavioral segments. Second, random forest classifiers were used to validate cluster coherence and quantify feature importance. Third, the Apriori algorithm was used to extract association rules that capture recurrent material-mixing behaviors. The results revealed municipality-specific segmentation structures (Tequixquiac: K = 6; Tlalpujahua: K = 3; Xalatlaco: K = 2), with material-specific disposal behaviors emerging as stronger segmentation drivers. Random forest classifiers validated cluster coherence with 100% accuracy, confirming that segments represent behaviorally distinct archetypes. The proposed framework converts raw behavioral data into actionable municipal visions. This approach focuses on finding diagnostic patterns instead of making predictions by utilizing machine-learning-driven MSWM research.
- Open Access
- 11 April 2023
- Published by Dirección de Bienes Públicos del Conocimiento, Universidad Autonoma del Estado de Mexico in CIENCIA ergo-sum
- Vol. 31
Estudio comparativo de tres Redes Neuronales Convolucionales pre-entrenadas en la detección de violencia contra las mujeres
This paper presents a performance comparison between three models of pre-trained CNN networks (VGG16, ResNet50, and MobileNet) in detecting physical violence against women in video. To carry out the classification of images that include physical violence against women and those that do not, 2 800 images (1 400 violent and 1 400 non-violent) were collected from a public dataset and subsequently divided into training (1 200 images), validation (1 000 images) and test (600 images). To evaluate their performance, accuracy values for each model were considered, positioning the MobileNet network as the best-performing classifier for this classification task with 89% accuracy.
- 11 April 2023
- Published by Dirección de Bienes Públicos del Conocimiento, Universidad Autonoma del Estado de Mexico in CIENCIA ergo-sum
- Vol. 31
Open AccessEstudio comparativo de tres Redes Neuronales Convolucionales pre-entrenadas en la detección de violencia contra las mujeres
This paper presents a performance comparison between three models of pre-trained CNN networks (VGG16, ResNet50, and MobileNet) in detecting physical violence against women in video. To carry out the classification of images that include physical violence against women and those that do not, 2 800 images (1 400 violent and 1 400 non-violent) were collected from a public dataset and subsequently divided into training (1 200 images), validation (1 000 images) and test (600 images). To evaluate their performance, accuracy values for each model were considered, positioning the MobileNet network as the best-performing classifier for this classification task with 89% accuracy.