Estimating semantic networks of groups and individuals from fluency data

Area of Science:

• Cognitive Psychology
• Computational Neuroscience
• Network Science

Background:

• Associative semantic networks are a classic model of knowledge representation in the mind.
• Estimating these semantic networks from data remains a challenge due to a lack of consensus on methodology.

Purpose of the Study:

• To propose and evaluate a novel method, U-INVITE, for estimating semantic networks from semantic fluency data.
• To compare U-INVITE against existing network estimation techniques using simulations and human judgment data.

Main Methods:

• Developed U-INVITE based on a censored random walk model of memory retrieval.
• Compared U-INVITE with other semantic network estimation methods using simulation studies.
• Assessed the psychological validity of estimated networks through human similarity judgments.

Main Results:

• U-INVITE demonstrated low error rates in recovering semantic networks with moderate amounts of semantic fluency data.
• U-INVITE is derived from a psychologically plausible process model and is a consistent estimator.
• Comparative analysis revealed varying merits among different network estimation techniques.

Conclusions:

• U-INVITE offers a robust and psychologically grounded method for semantic network estimation.
• The study provides a framework for selecting appropriate network estimation methods.
• Open-source code is provided to facilitate the application of these techniques.