The model showed the garden paths deserved further investigation

Andrew Perun, Mandy Osuji, Rishika Veeramachaneni December 2024

Data Preprocessing:

Download word frequency data, surprisal data from Arehalli et al.'s supertag-trained models (2023), and reading times (ClassicGardenPathSet.csv and Fillers.csv) into main directory defined at "base_path" path. Run the code in LING_380_Preprocess.ipynb to produce fmerged_mod.csv and merged_mod.csv files, to be used in LME experiments.

Fitting LME Models:

Define a base path for experiment results in Project_Final.ipynb at "base". More pre-processing to create lagged datasets is necessary before creating new combined datasets and producing reading time predictions.