Computer Science
3 years ago

New Information Criterion Outperforms Traditional Models in System Identification.

Parameter Magnitude-Based Information Criterion in Identification of Discrete-Time Dynamic System / Md Fahmi Abd Samad and Abdul Rahman Mohd Nasir
Fahmi Abd Samad, Abdul Rahman Mohd Nasir

Paper Summary

A new information criterion called PMIC2 was introduced to help choose the best model for dynamic systems. The researchers compared PMIC2 with other popular criteria like AIC and BIC, using computer simulations. The results showed that PMIC2 was better at selecting the most accurate and simple model structure compared to the other criteria.

Related papers

Theme background image
Mathematics
8 years ago

New information criteria revolutionize model selection in linear regression.

Regularization Methods
Bayesian Networks
Model-Based Clustering
Theme background image
Engineering
5 years ago

New autoregression criterion adapts to data, outperforming traditional model selection methods.

System Identification
Blind Source Separation
Neural Networks
Theme background image
Engineering
7 years ago

New Noncentral Information Criterion revolutionizes model selection in linear regression!

Process Fault Detection
Blind Source Separation
Neural Networks
1 view