Vanessa Villazón León

[Vanessa Villazón León @villazon_leon]

Hello @LatEnvChemPSE community, I am happy to share our poster "Machine learning algorithms evaluation for predicting the triple point temperature of pure compounds" #LatinEnvChemPSE #PSEModSimulatControl

[Vanessa Villazón León @villazon_leon]

@OctavioDMA @LatEnvChemPSE Sin embargo, la introducción de la temperatura y presión crítica como variables de entrada al modelo, no contribuyeron a mejorar el desempeño del mismo.

[Vanessa Villazón León @villazon_leon]

Hello @LatEnvChemPSE community, I am happy to share our poster "Machine learning algorithms evaluation for predicting the triple point temperature of pure compounds" #LatinEnvChemPSE #PSEModSimulatControl

[Vanessa Villazón León @villazon_leon]

@OctavioDMA @LatEnvChemPSE Sería interesante evaluar el desempeño del modelo si se introdujeran variables adicionales como el factor acéntrico y la temperatura de ebullición normal. Es algo que no he evaluado.

[Vanessa Villazón León @villazon_leon]

@OctavioDMA @LatEnvChemPSE Sí, corresponden al modelo M1, que fue el que mostró mejor desempeño.

[Vanessa Villazón León @villazon_leon]

@OctavioDMA @LatEnvChemPSE Muchas gracias Octavio!

[Vanessa Villazón León @villazon_leon]

@RFonsPer @LatEnvChemPSE By combining multiple models, it reduces both bias and variance. This helps to create a more robust model that can generalize better to new data. This algorithm is less affected by outliers compared to other ML algorithms, minimizing the impact of atypical data points.

[Vanessa Villazón León @villazon_leon]

@RFonsPer @LatEnvChemPSE Gradient boosting algorithm combines different weak models (usually decision trees) to create a strong one. Each subsequent model aims to rectify the errors of the previous one, improving the prediction accuracy.

[Vanessa Villazón León @villazon_leon]

@Chalitogarnet @LatEnvChemPSE Hola, muchas gracias. Para el desarrollo de modelos termodinámicos he estado usando Matlab. Sin embargo, en este Review se abordan diferentes trabajos que utilizan distintos lenguajes de programación como python, entre otros.

[Vanessa Villazón León @villazon_leon]

Hola, les comparto nuestro trabajo titulado "Análisis de modelos de contribución de grupos para la estimación de propiedades termodinámicas de líquidos iónicos" @LatEnvChemPSE #PSEModSimulatControl

[Vanessa Villazón León @villazon_leon]

@AlanAguirreSoto @RoySocChem Thank you!!!

[Vanessa Villazón León @villazon_leon]

Hello #RSCPoster community, I'm excited to share our poster entitled "Artificial neural networks and group contribution approaches for the estimation of heat capacities of pure compounds" @RoySocChem #RSCEng.

[Vanessa Villazón León @villazon_leon]

@RichBourne1981 @RoySocChem 3. We collected the experimental database using the National Institute of Standards and Technology (NIST).

[Vanessa Villazón León @villazon_leon]

@RichBourne1981 @RoySocChem 2. https://t.co/xzHmitYBpe
https://t.co/Ebq6QI5GWC

[Vanessa Villazón León @villazon_leon]

@RichBourne1981 @RoySocChem 1. Thanks for your comments. Low molecular weight compounds often don't follow the general trend of the homologous series of a specific chemical family. This behavior has been explained by Nannoolal et al. (2007) and other authors in the estimation of different properties.

[Vanessa Villazón León @villazon_leon]

@ZHAIWeiNUS @RoySocChem Thank you for your comment. Figure 3 shows the overall database error, including the training, validation and test sets.

[Vanessa Villazón León @villazon_leon]

@RonaldMarq @RoySocChem Thank you, we should evaluate different machine learning algorithms in order to compare their performance with those obtained using ANN methods. In the next steps of our work we will do this.

[Vanessa Villazón León @villazon_leon]

@WielgusWW @RoySocChem Thank you!! We are going to evaluate the models performance using new experimental data.

[Vanessa Villazón León @villazon_leon]

@EliasMH14 @LatEnvChemPSE Thank you for the support and suggestions on our work.

[Vanessa Villazón León @villazon_leon]

Hello @LatEnvChemPSE community, it's a pleasure to share my poster entitled "Estimation of heat capacity of pure compounds using group contribution and artificial neural networks" #PSEFundComp

[Vanessa Villazón León @villazon_leon]

@EmanuelContC @LatEnvChemPSE Hello, there is no relation between the molecules used for the model development. We used both organic and inorganic compounds from different chemical families.