SteveW
Are there strategies for avoiding overfitting a model to the most common state? I've got a binary target variable and 91% precision on the common state and 31% on the uncommon state. This roughly follows my distribution of observations (3:1).I would like to achieve a better balance, even at the cost of overall model fit. Stratification is not affecting results (I think it affects only network complexity not parameter estimation?thanks
Quote 0 0
Dan
Indeed, the stratification just allows to change the marginal distributions of the selected set of nodes in order to get a more complex model. However, at the end of the learning algorithms, the parameters are estimated on the unstratified data. We will add an option to allow the estimation of the parameters on the stratified dataset. In the meantime, you can try to sample your dataset, or even better, associate a higher weight with the data corresponding to the less likely state.Hope this helps
Quote 0 0
Dan
Hi Steve,Just to let you know that we just released version 6.0.5 that comes with the option to use the stratification for the parameter estimation.Best,Dan
Quote 0 0

Add a Website Forum to your website.