#PVL_MLEfitting_IGT.r
#This program does the Prospect Valence Learning (PVL), Baseline, & Random Models for the IGT
#PVL is the 4-parameter interference version
#This program automatically handles data files where the number of trials varies by subject.
#It can also handle deck depletion if you modify the cardsperdeck variable.
#This program will only model up to the trial where one deck runs out.
#Originally programmed by Anthony Bishara for the EVL model (Busemeyer & Stout, 2002).
#Visit his homepage @ http://bisharaa.people.cofc.edu/
#
#Modified by Woo-Young Ahn for the PVL model (Ahn, Busemeyer, Wagenmakers, & Stout, 2008)
#This code estimates 4 free parameters of the PVL model using maximum likelihood estimation
#You can try two different learning rules and two different versions of the Luce choice rule
#See Ahn et al. (2008) for more details.
# Two learning rules: 1. delta-rule, 2: decay-reinforcement (RI) learning rule
# Two choice rules: 1. trial-dependent , 2: trial-independent
# Bounds of each parameter (different from settings in Ahn et al. (2008) for c and lambda)
#   A (learning rate): 0.0001 to 0.9999
#   alpha: from 0.01 to 1
#   lambda: from 0.01 to 5
#   -5 < c (sensitivity) < 5 for trial-dependent choice rule
#   0 < c (sensitivity) < 5 for trial-independent choice rule
#For questions, you can email Woo-Young Ahn (ahnw@indiana.edu)

#Instructions:
#1) Download and install R from http://www.r-project.org/.  When you get to the options, choose "base" and choose the executable file (e.g. R-2.3.1-win32.exe)
#2) Run R.
#3) In the program, click File, then Change Dir¡¦, then choose the directory your data is in.
#4) Use tab-delimated text format for the data. Columns=trial, deck, wins, losses, and subject.
#   Extra columns after the subject column are okay.
#   Don't use column labels.
#4) Modify the green sections below for your dataset.
#5) Select all the text in this document (ctrl+a), then copy and paste into R
#   or just source the file from R.
#6) The output file is labelled "PVL_Results_" plus submodel names and the name of the data file.