Estimate Networks Demo

This demo estimates semantic networks from fluency data using a variety of modeling techniques implemented in SNAFU. It provides a comparative view of different network construction algorithms commonly used in semantic network research.

—

Overview of Workflow:

  1. Load animal fluency data for a specific group (Experiment1), applying spell correction and flattening the data.

  2. Define fit parameters using the Fitinfo object (mainly for Conceptual Network).

  3. Estimate semantic networks using five different methods: - Naive Random Walk (NRW) – Random walk transition probabilities - Conceptual Network (CN) – Co-occurrence-based estimation from Goni et al. (2011) - Pathfinder Network (PF) – Based on distance metrics - Correlation-Based Network (CBN) – Based on word correlation - First-Edge Network (FE) – Based on order of item appearance

  4. Save each network’s edge list as a .csv file for further visualization or analysis.

Functions Used:

  • snafu.load_fluency_data

  • snafu.Fitinfo

  • snafu.naiveRandomWalk

  • snafu.conceptualNetwork

  • snafu.pathfinder

  • snafu.correlationBasedNetwork

  • snafu.firstEdge

  • snafu.write_graph

Output Files:

Each file contains an edge list in CSV format:

  • nrw_graph.csv

  • cn_graph.csv

  • pf_graph.csv

  • cbn_graph.csv

  • fe_graph.csv

All files are saved in the demos_data/ directory and labeled by group.