 Hi, today I want to present this work called residual RWA, detecting relevant variables using relative wave analysis with recent residualization. This work is a joint work with my colleague Carlos Cascar. We are from the University of Costa Rica in the Simba. I want to ask, what is residual wave analysis? This is a technique to detect the contribution of variables or effects in a model. The idea that it projects the original space data to an orthonormal space where it splits the weights of each variable. In fact, one of the biggest results is that some of the weights are approximately the models are square. So, our methodology follows three steps. First, we select the best model, giving a control of three bits and interaction terms. Then we visualize all interaction terms, where we survey the interaction effects from the main ones. And third, we apply a relative wave analysis where we detect which variables are the most relevant. In the first step, we use a location information criterion to select the best model between a simple model or lower model that only has a control and fix terms and an upper model that has a control, fix, free and interaction terms. The functions FI could be linear or a restricted cubic response to add a nonlinear effect in each term. Then we visualize all the interaction terms, the visualization basically removes all the main effects for each interaction. Here, we substitute each interaction by a third residualized version once we make all the equations appropriately. Finally, we apply a relative wave analysis to the design matrix with the main and interaction effects after residualization. Basically, we project the design matrix to our normal space. Then we estimate fully the standardized coefficients in this space. Then we estimate the relative importance of those coefficients for this configuration. And finally, we transform the coefficients into the original space to have the relative points. This procedure gives the important weights of each term, but in the reduced model. The package, the SIDL RWA, has the main function, which including the response control fix free variables and also the kind of farming that we want to apply. We can estimate all this procedure as we explain. In fact, using a simple Ishigami model, which contains free and interaction terms that are relevant to the model. We obtain a very accurate result according to the Ishigami model. Of course, we need a lot of more simulations, but for our other obtained, we can present to you. Finally, I want to present the paper that we work with all the methodology in this package. It's already in archive and is submitted to be published. We hope this year. Also, the package called the SIDL RWA is already in GitHub, and we hope to submit it to plan in the next weeks. Thank you very much for your attention.