Choose ANN parameters with a validating test instead of typical statistical criteria such as AIC or BIC¶

Strategy · All strategies
Evidence · no claims cited

Description¶

The article recommends selecting the appropriate parameters for the ANNs using a validating test rather than typical statistical criteria such as AIC or BIC. The authors present this as a distinctive feature of their approach within cognitive diagnostic modelling. It supports the goal of achieving robust and accurate classification of latent attribute profiles.

Design Implications¶

Context¶

Requirements¶

• Requires a validating test against which ANN parameter choices can be evaluated

Constraints¶

Target Learners¶

• students assessed with diagnostic classification assessments

Target Learning Goals¶

• accurate classification of students' latent attribute profiles

Related Strategies¶

Examples¶

-

Key Sources¶

• Kang Xue, Laine Bradshaw. (2021). A semi-supervised learning-based diagnostic classification method using artificial neural networks. Frontiers in Psychology, 11, 3992. https://www.frontiersin.org/journals/psychology