Browsing by Author "Ruz, Gonzalo A."

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    A RUL Estimation System from Clustered Run-to-Failure Degradation Signals
    (2022) Cho, Anthony D.; Carrasco Schmidt, Rodrigo Arnaldo; Ruz, Gonzalo A.
  • No Thumbnail Available
    Item
    Early successional patterns of bacterial communities in soil microcosms reveal changes in bacterial community composition and network architecture, depending on the successional condition
    (2017) Rodriguez-Valdecantos, Gustavo; Manzano, Marlene; Sanchez, Raimundo; Urbina, Felipe; Hengst, Martha B.; Antonio Lardies, Marco; Ruz, Gonzalo A.; Gonzalez, Bernardo
    Soil ecosystem dynamics are influenced by the composition of bacterial communities and environmental conditions. A common approach to study bacterial successional dynamics is to survey the trajectories and patterns that follow bacterial community assemblages; however early successional stages have received little attention. To elucidate how soil type and chemical amendments influence both the trajectories that follow early compositional changes and the architecture of the community bacterial networks in soil bacterial succession, a time series experiment of soil microcosm experiments was performed. Soil bacterial communities were initially perturbed by dilution and subsequently subjected to three amendments: application of the pesticide 2,4-dichlorophenoxyacetic acid, as a pesticide-amended succession; application of cycloheximide, an inhibitor affecting primarily eukaryotic microorganisms, as a eukaryotic-inhibition bacterial succession; or application of sterile water as a non-perturbed control. Terminal restriction fragment length polymorphism (T-RFLP) analysis of the 16S rRNA gene isolated from soil microcosms was used to generate bacterial relative abundance datasets. Bray-Curtis similarity and beta diversity partition-based methods were applied to identify the trajectories that follow changes in bacterial community composition. Results demonstrated that bacterial communities exposed to these three conditions rapidly differentiated from the starting point (less than 12 h), followed different compositional change trajectories depending on the treatment, and quickly converged to a state similar to the initial community (48-72 h). Network inference analysis was applied using a generalized Lotka-Volterra model to provide an overview of bacterial OTU interactions and to follow the changes in bacterial community networks. This analysis revealed that antagonistic interactions increased when eukaryotes were inhibited, whereas cooperative interactions increased under pesticide influence. Moreover, central OTUs from soil bacterial community networks were also persistent OTUs, thus confirming the existence of a core bacterial community and that these same OTUs could plastically interact according to the perturbation type to quickly stabilize bacterial communities undergoing succession.
  • No Thumbnail Available
    Item
    Modeling Recidivism through Bayesian Regression Models and Deep Neural Networks
    (2021) de la Cruz, Rolando; Padilla, Oslando; Valle, Mauricio A.; Ruz, Gonzalo A.
    This study aims to analyze and explore criminal recidivism with different modeling strategies: one based on an explanation of the phenomenon and another based on a prediction task. We compared three common statistical approaches for modeling recidivism: the logistic regression model, the Cox regression model, and the cure rate model. The parameters of these models were estimated from a Bayesian point of view. Additionally, for prediction purposes, we compared the Cox proportional model, a random survival forest, and a deep neural network. To conduct this study, we used a real dataset that corresponds to a cohort of individuals which consisted of men convicted of sexual crimes against women in 1973 in England and Wales. The results show that the logistic regression model tends to give more precise estimations of the probabilities of recidivism both globally and with the subgroups considered, but at the expense of running a model for each moment of the time that is of interest. The cure rate model with a relatively simple distribution, such as Weibull, provides acceptable estimations, and these tend to be better with longer follow-up periods. The Cox regression model can provide the most biased estimations with certain subgroups. The prediction results show the deep neural network's superiority compared to the Cox proportional model and the random survival forest.