Generator of Synthetic Data for the Evaluation of Context-Aware Recommendation Systems


DataGenCARS is a complete Java-based synthetic dataset generator for the evaluation of Context-Aware Recommendation Systems (CARS).  The generator presents a high flexibility in the obtaining of datasets in several scenarios of items recommendation (with context or without context).

Some key features of Data-GenCARS are listed below:

- Flexible de finition of user schemas, item schemas and context schemas.

- Flexible de nfition of user profi les.

- Realistic generation of ratings and attributes of items.

- The automatic mapping between item schemas and Java classes.

- The possibility to mix real and synthetic datasets and of replicating existing real datasets.

User Manual:

DataGenCARS Quick user guide

DataGenCARS Quick user guide-1.1 beta

TestDataGenCARS-1.2 beta

Java documentation:  The Java documentation generated by DataGenCARS can be downloaded here.

Download API:


datagencars-1.1 (beta)

datagencars-1.2 (beta)

Main Related Publications:


M.C. Rodríguez-Hernández, S. Ilarri, R. Hermoso and R. Trillo-Lado, “DataGenCARS: A generator of synthetic data for the evaluation of context-aware recommendation systems“,  Pervasive and Mobile Computing, ISSN 1574-1192. (DOI: 10.1016/j.pmcj.2016.09.020)

International Conferences

M.C. Rodríguez-Hernández, S. Ilarri, R. Hermoso, and R. Trillo-Lado. “Towards trajectory-based recommendations in museums: Evaluation of strategies using mixed synthetic and real data“, 8th International Conference on Emerg-ing Ubiquitous Systems and Pervasive Networks (EUSPN), volume 113, pp. 234-239. September 2017. ISSN 1877-0509. DOI 10.1016/j.procs.2017.08.355. Procedia Computer Science.