Matrix.ita Software.som

: 10,000 flights JFK to LAX. Combining them with 5,000 connecting flights yields 50M itineraries.

from minisom import MiniSom import numpy as np fare_matrix = np.random.rand(1000, 10) # 1000 routes, 10 fare classes som = MiniSom(5, 5, 10, sigma=0.5, learning_rate=0.5) som.train_random(fare_matrix, 100) The winning node is your optimal fare cluster (the "ITO" - Itinerary Optimization) print(som.winner(fare_matrix[0])) matrix.ita software.som

Use the opensky-network API or OAG’s data feed to collect flight schedules. Store them in a 3D NumPy array. : 10,000 flights JFK to LAX