Amazon Commerce Reviews for Multi-Class Text Classification

Amazon commerce reviews dataset containing 2000 rows, stratified across 10 classes. The data was sourced from OpenML, a platform for sharing machine learning datasets. Specific details about the author, collection method, and temporal coverage are not provided in the input metadata.

Use Cases

• Classify review sentiment into one of 10 categories using text features.
• Analyze the relationship between review text and its assigned class label for model interpretability.
• Train a text classifier to predict the review class from the 100 provided features.
• Benchmark multi-class classification algorithms on a stratified sample of 2000 reviews.

Strengths

• Dataset contains 2000 rows for model training and evaluation.
• Data is stratified across 10 target classes, promoting balanced learning.
• Includes 100 columns or features per row for potential analysis.

Limitations

• Sample size of 2000 rows may be insufficient for training complex deep learning models.
• Lack of column descriptions makes feature interpretation and selection challenging.
• Unknown temporal range and collection method limit analysis of trends over time.

Provenance

Source

OpenML