Welcome to this Github for my Master Thesis in Data Science (with the same title).

Here, you will find

• The implementations of all the models used (for each experiment),
• The raw results for all the experiments,
• The evaluation codes and raw evaluations.

It it ordered by research question, test, category, and number of reviews (nr_rev).

If you want to run the code yourself, you should first download the data you want to use here.

After downloading the data, store it in a folder called 'Data'. Check whether the files are still zipped and consider unzipping them. Large files take a long time, it is recommended to use model_RQ1.py to first separate large jsons into smaller csvs. Put all these csvs in a folder within 'Data' called 'csvs_{category}' or, if meta data, 'csvs_meta_{category}'.

If all goes well, you can either slightly adjust the .py file to loop through the different categories and nr_revs themselves or you can use the terminal.

After opening your terminal in the same folder as your 'Data' and codes location an example run in the terminal would be:

RQ1: python3 model_RQ2.py {category} {nr_rev} for instance python3 model_RQ1.py Office_Products 50

RQ2: python3 model_RQ1.py {test} {category} for instance python3 model_RQ1.py test3 Home_and_Kitchen

RQ3b: python3 model_RQ2.py {time-window} {category} for instance python3 model_RQ1.py 90 Automotive

Note: The loading in can take quite a while so do not be alarmed.