MovieLens Ratings Injected with Shilling Attacks for Robustness Testing

A modified version of the MovieLens dataset, created for testing the robustness of recommender systems. The raw description indicates it has been injected with shilling attacks, which are artificial profiles designed to manipulate recommendation outputs. The dataset's author, organization, and specific version details are unknown.

Use Cases

• Testing recommender system robustness against profile injection attacks based on the described shilling attacks.
• Benchmarking anomaly detection methods for identifying malicious user profiles in rating data.
• Studying the impact of data poisoning on the accuracy and fairness of movie recommendation models.

Strengths

• Designed for a specific, high-value research task: testing recommender system robustness against adversarial attacks.
• Builds upon the well-known and widely used MovieLens dataset, providing a familiar baseline.

Limitations

• Row count, column definitions, and file formats are unknown, which limits suitability assessment.
• Description metadata is limited; actual data quality and attack implementation require manual inspection after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

Kaggle

Collection Method

Likely derived from the original MovieLens dataset with injected adversarial profiles.