Code for explorations on probabilistic learning and decision-making tasks; for learning purposes
To explore the pipeline and the role of each function, see: single_ptp_demo.m: a demo for 1) running a single simulation and 2) perform model fitting for the virtual participant generated
compare_PL_RW1lrs_1: script for exploring how WSLS/accuracy/task earning are influenced by learning rate and inverse temperature for a simple Rescorla-Wagner (RW) learning model with Softmax action selection, under different task structures (e.g., stable versus volatile). Also touches on comparing between RW model with or without updating the value of the unchosen option.
Some of the simulation results are presented in Eddie_PaLPresentation_9Feb_GitHubVersion.pdf
Other functions: response_model_illu: produce plots explaining how inverse temperature and lapse rate influences the relationship between expected values and probability of choice