Module model_selection 

Model Selection Guidance

This module provides intelligent model selection and recommendation capabilities to help users choose the most appropriate embedding model for their specific knowledge graph and use case.

Features

• Automatic Model Recommendation: Based on dataset characteristics
• Model Comparison: Compare multiple models on the same dataset
• Performance Profiling: Benchmark model performance
• Resource Requirements: Estimate memory and compute needs
• Use Case Matching: Recommend models for specific applications

Example Usage

use oxirs_embed::model_selection::{ModelSelector, DatasetCharacteristics, UseCaseType};

// Define dataset characteristics
let characteristics = DatasetCharacteristics {
    num_entities: 10000,
    num_relations: 50,
    num_triples: 50000,
    avg_degree: 5.0,
    is_sparse: false,
    has_hierarchies: true,
    has_complex_relations: true,
    domain: Some("biomedical".to_string()),
};

// Get model recommendations
let selector = ModelSelector::new();
let recommendations = selector.recommend_models(&characteristics, UseCaseType::LinkPrediction)?;

for rec in recommendations {
    println!("Model: {}, Score: {:.2}, Reason: {}",
             rec.model_type, rec.suitability_score, rec.reasoning);
}

Structs

DatasetCharacteristics

Dataset characteristics for model selection

ModelComparison

Comparison result for multiple models

ModelComparisonEntry

Entry in model comparison

ModelRecommendation

Model recommendation with reasoning

ModelSelector

Model selector for intelligent recommendation

Enums

MemoryRequirement

Memory requirement estimate

ModelType

Available embedding model types

TrainingTime

Training time estimate

UseCaseType

Type of use case for model selection