WebPredictionResult class

Represents the result of a web-based mosquito species prediction.

This model contains prediction results from remote classification services, including species identification, confidence scores, probability distributions across all possible species, and associated metadata. It's used when classification is performed on external servers rather than locally on the device.

The prediction result includes comprehensive probability data for transparency and supports both single-prediction and multi-class classification scenarios.

Usage Example

// Typically received from API response
final result = WebPredictionResult(
  id: 'pred_123456',
  scientificName: 'Aedes aegypti',
  probabilities: {
    'Aedes aegypti': 0.87,
    'Aedes albopictus': 0.08,
    'Culex pipiens': 0.03,
    'Anopheles gambiae': 0.02,
  },
  modelId: 'culico-net-cls-v1',
  confidence: 0.87,
  imageUrlSpecies: 'https://api.culicidaelab.org/species/aedes_aegypti.jpg',
);

print('Top prediction: ${result.scientificName}');
print('Confidence: ${result.confidencePercentage}');
print('Alternative species: ${result.alternativeSpecies}');

Server Integration

This model is designed to work with the CulicidaeLab server API:

• Receives predictions from /predict endpoint
• Includes full probability distribution for transparency
• Supports model versioning and tracking
• Provides species image URLs for display

Comparison with Local Classification

Unlike ClassificationResult which is used for local PyTorch Lite predictions, this model represents server-side classification results with additional features like probability distributions and remote image URLs.

See also:

• ClassificationResult for local device-based predictions
• Observation for storing prediction results as observations
• CulicidaeLab server API documentation