Smart Seed Selection
Smart Seed
Selection
Smart Seed Selection
Large № of farmers
Soil analysis
Weather
Genomics
Satellite image processing
Machine learning
Big Data analytics
Data
Data Fusion by Matrix Factorisation
• Sparse dataset
• Matrix decompositioninto latent matrices
• Imputation of missing values
• Yield prediction using machine learning
Seed Selection using Portfolio Optimisation
• Where to invest?
• Markowitz’s Modern Portfolio Theory
• Investments = seed varieties
• Source of risk: weather
Multi-Objective Evolutionary Algorithms
• Profit
• Risk
• Resistance
• Use of fertilisers
• Use of pesticides
• Crop rotation
• Planting date
Smart Seed Selection
• Additional investment = 0
• Sustainability
• Growing system
Thank you!
bigdata.cgiar.org