Eye Movements: Relevance Inference from Eye Tracking Data

Jarkko Salojarvi et al. from Helsinki University of Technology published this dataset in 2005. It contains pre-computed feature vectors derived from eye movement trajectories, designed for a classification task. The data is structured around assignments where a question is followed by ten sentences, each labeled as correct, relevant, or irrelevant.

Use Cases

• Train classifiers to infer relevance from eye movement patterns based on pre-computed feature vectors.
• Study human information processing and reading behavior based on eye tracking data.
• Benchmark models for predicting sentence relevance in a question-answering context.
• Analyze the relationship between visual attention and textual comprehension.

Strengths

• Data is preprocessed and includes pre-computed feature vectors, reducing initial feature engineering effort.
• Provides a clear, multi-class labeling scheme (Correct, Relevant, Irrelevant) for a classification task.
• Originates from a published academic report, suggesting a research-grade foundation.

Limitations

• Row count, column definitions, and file formats are unknown, limiting suitability assessment.
• Last update date is unknown; the dataset is from 2005, so freshness is unverified.
• Column-level documentation is absent; field semantics must be inferred after download.

Provenance

Source

Helsinki University of Technology, Publications in Computer and Information Science, Report A82.

Collection Method

Likely gathered from controlled experiments involving eye tracking during reading tasks.

Time Range

Publication date is 2005; specific data collection period is unknown.

Freshness

Last updated date is unknown; data originates from a 2005 publication.

Geography

Geography is not specified in the provided description.