RB-CRS

"Generation-based vs. Retrieval-based Conversational Recommendation: A User-Centric Comparison"

We have two main directories at the root level.

1. The first named ‘RB-CRS Code’ contains the complete Python source code for our own developed retrieval-based recommender system.

2. The second directory, named ‘Data’ includes the data and results for the online study reported in the manuscript.

Before running the scripts in the directory named 'RB-CRS Code', please download the ratings data,i.e., 'ratings_latest.csv' file, from this link. Add this file to the path ../CRB-CRS/data/recommenders_item_data/

You may also download ratings data directly from the MovieLens website

Citation

inproceedings{10.1145/3460231.3475942,
author = {Manzoor, Ahtsham and Jannach, Dietmar},
title = {Generation-Based vs. Retrieval-Based Conversational Recommendation: A User-Centric Comparison},
year = {2021},
isbn = {9781450384582},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3460231.3475942},
doi = {10.1145/3460231.3475942},
booktitle = {Fifteenth ACM Conference on Recommender Systems},
pages = {515–520},
numpages = {6},
keywords = {End-to-end Learning, Evaluation, Conversational Recommendation, Retrieval, Language Generation},
location = {Amsterdam, Netherlands},
series = {RecSys '21}
}