Soybean Disease Classification with 35 Plant and Weather Attributes

19 soybean disease classes are predicted from 35 categorical attributes describing plant characteristics, weather conditions, and observed damage. The data, originally from UCI, includes attributes like date, precipitation, hail, leaf spots, and stem cankers, with values encoded numerically and 'dna' or '?' for missing information. Authors R.S. Michalski and R.L. Chilausky compiled this dataset for machine learning research.

Use Cases

• Benchmarking multi-class classification algorithms based on the 19 disease categories.
• Studying feature importance for disease diagnosis based on the 35 categorical plant and environmental attributes.
• Exploring methods for handling categorical data and missing values encoded as '?' and 'dna'.
• Training models for automated soybean disease identification in agricultural monitoring systems.

Strengths

• Includes 35 distinct categorical features covering plant morphology, symptoms, and environmental factors.
• Provides a defined multi-class prediction task with 19 disease categories.
• Explicit documentation of encoding schemes for values and missing data ('dna', '?').

Limitations

• Row count and file format are unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.
• The description notes that the last four of the 19 classes have very few examples, which could impact model performance.

Provenance

Source

UCI Machine Learning Repository, authors R.S. Michalski and R.L. Chilausky.

Collection Method

null

Time Range

null

Freshness

null

Geography

null