need to explore and choose ways to do model selection

domenicodigangi / DynWGraphsPaper

Code for analysis of paper about dynmical sparse weighted graphs

MIT License

0 stars 1 forks source link

need to explore and choose ways to do model selection #9

Open domenicodigangi opened 3 years ago

domenicodigangi commented 3 years ago

For example, for a given regressor, how should we decide wether beta is 0 or not? how should we discriminate between a model with one or multiple betas.

[ ] Repeat literature search for generic GLM models (are there alternatives to AIC and BIC?)
[ ] Can we test wether a criterium is practically justified by numerical simulations? For example can I simulate a model with single beta and check if AIC and BIC select it over a version with multiple beta?

domenicodigangi commented 3 years ago

LOGISTIC NETWORK REGRESSION FOR SCALABLE ANALYSIS OF NETWORKS WITH JOINT EDGE/VERTEX DYNAMICS Zack W. Almquist* and Carter T. Butts - Compute distributions of Graph Level Indices on , simulated out of sample (forecasted) networks and compare them with the observed ones . Close to the idea of checking for the pox to forecast network statistics