File size: 17,425 Bytes
8986ff6 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 |
"""
Layer 2: Upstream Trade Flow Modeling
====================================
Models disruption to global trade networks using graph neural networks
and trade flow reallocation predictions.
"""
from typing import Dict, List, Any, Optional, Tuple, Union
import numpy as np
import pandas as pd
from dataclasses import dataclass
import networkx as nx
from sklearn.preprocessing import StandardScaler
from sklearn.ensemble import RandomForestRegressor
from ..config.layer_configs import TradeFlowConfig
@dataclass
class TradeFlowPrediction:
"""Trade flow impact prediction results"""
affected_routes: List[Dict[str, Any]]
reallocation_matrix: np.ndarray
supply_chain_disruption: Dict[str, float]
alternative_sources: Dict[str, List[str]]
cost_increases: Dict[str, float]
time_delays: Dict[str, float]
class TradeFlowLayer:
"""
Layer 2: Trade Flow Modeling
Predicts supply chain disruption and trade flow reallocation
using graph-based representation of global trade networks.
"""
def __init__(self, config: Union[TradeFlowConfig, Dict[str, Any]]):
"""Initialize trade flow layer"""
# Handle both TradeFlowConfig and dict types
if hasattr(config, "__dict__") and not isinstance(config, dict):
# Convert TradeFlowConfig or similar object to dict
self.config: Dict[str, Any] = config.__dict__
else:
self.config: Dict[str, Any] = config
# Trade network graph
self.trade_graph = nx.DiGraph()
self.trade_data = None
# ML models
self.flow_predictor = RandomForestRegressor(n_estimators=100, random_state=42)
self.cost_predictor = RandomForestRegressor(n_estimators=100, random_state=42)
# Feature processing
self.feature_scaler = StandardScaler()
self.is_fitted = False
# Trade route mappings
self.country_to_index = {}
self.index_to_country = {}
self.hs_code_mappings = {}
def fit(self, trade_data: Optional[pd.DataFrame]) -> "TradeFlowLayer":
"""
Train trade flow models on historical data
Args:
trade_data: DataFrame with columns:
- origin_country: Origin country code
- destination_country: Destination country code
- hs_code: Harmonized System product code
- trade_value: Trade value in USD
- year: Year of trade
- transport_cost: Transportation cost
- lead_time: Lead time in days
"""
if trade_data is None or trade_data.empty:
self._initialize_default_network()
self.is_fitted = True
return self
self.trade_data = trade_data.copy()
# Build trade network graph
self._build_trade_graph(trade_data)
# Prepare training features
features, targets = self._prepare_training_data(trade_data)
if len(features) > 0:
# Fit feature scaler
self.feature_scaler.fit(features)
features_scaled = self.feature_scaler.transform(features)
# Train flow prediction model
flow_targets = [t["flow_change"] for t in targets]
self.flow_predictor.fit(features_scaled, flow_targets)
# Train cost prediction model
cost_targets = [t["cost_change"] for t in targets]
self.cost_predictor.fit(features_scaled, cost_targets)
self.is_fitted = True
return self
def _initialize_default_network(self):
"""Initialize default trade network for demo purposes"""
# Major trading countries
major_countries = ["US", "CN", "DE", "JP", "UK", "FR", "NL", "SG", "KR", "IN"]
# Create country mappings
self.country_to_index = {
country: i for i, country in enumerate(major_countries)
}
self.index_to_country = {
i: country for i, country in enumerate(major_countries)
}
# Add nodes to graph
for country in major_countries:
self.trade_graph.add_node(country, trade_volume=1000000)
# Add default trade relationships (simplified)
default_routes = [
("CN", "US", 500000),
("US", "CN", 400000),
("CN", "DE", 300000),
("DE", "CN", 250000),
("US", "DE", 200000),
("DE", "US", 180000),
("CN", "JP", 250000),
("JP", "CN", 220000),
("SG", "CN", 150000),
("CN", "SG", 160000),
("SG", "US", 100000),
("US", "SG", 90000),
]
for origin, dest, volume in default_routes:
self.trade_graph.add_edge(
origin, dest, trade_volume=volume, transport_cost=0.05, lead_time=14
)
def _build_trade_graph(self, trade_data: pd.DataFrame):
"""Build trade network graph from historical data"""
# Get unique countries
countries = list(
set(
trade_data["origin_country"].unique().tolist()
+ trade_data["destination_country"].unique().tolist()
)
)
# Create country mappings
self.country_to_index = {country: i for i, country in enumerate(countries)}
self.index_to_country = {i: country for i, country in enumerate(countries)}
# Add nodes
for country in countries:
country_volume = trade_data[
(trade_data["origin_country"] == country)
| (trade_data["destination_country"] == country)
]["trade_value"].sum()
self.trade_graph.add_node(country, trade_volume=country_volume)
# Add edges (trade routes)
route_aggregates = (
trade_data.groupby(["origin_country", "destination_country"])
.agg({"trade_value": "sum", "transport_cost": "mean", "lead_time": "mean"})
.reset_index()
)
for _, row in route_aggregates.iterrows():
self.trade_graph.add_edge(
row["origin_country"],
row["destination_country"],
trade_volume=row["trade_value"],
transport_cost=row["transport_cost"],
lead_time=row["lead_time"],
)
def _prepare_training_data(
self, trade_data: pd.DataFrame
) -> Tuple[List[List[float]], List[Dict[str, float]]]:
"""Prepare training features and targets"""
features = []
targets = []
# Simulate historical tariff shocks for training
# In practice, this would use actual historical policy data
sample_size = min(1000, len(trade_data))
sample_data = trade_data.sample(n=sample_size, random_state=42)
for _, row in sample_data.iterrows():
# Extract features
feature_vector = self._extract_route_features(
row["origin_country"], row["destination_country"], row["hs_code"]
)
# Simulate targets (would be actual historical impacts)
target = {
"flow_change": np.random.normal(0, 0.2), # ±20% flow change
"cost_change": np.random.uniform(0, 0.1), # 0-10% cost increase
}
features.append(feature_vector)
targets.append(target)
return features, targets
def _extract_route_features(
self, origin: str, destination: str, hs_code: str
) -> List[float]:
"""Extract features for a trade route"""
features = []
# Basic route features
if self.trade_graph.has_edge(origin, destination):
edge_data = self.trade_graph[origin][destination]
features.extend(
[
np.log1p(edge_data.get("trade_volume", 1)),
edge_data.get("transport_cost", 0.05),
edge_data.get("lead_time", 14),
]
)
else:
features.extend([0, 0.1, 30]) # Default values for missing routes
# Country-level features
origin_volume = (
self.trade_graph.nodes[origin].get("trade_volume", 0)
if origin in self.trade_graph
else 0
)
dest_volume = (
self.trade_graph.nodes[destination].get("trade_volume", 0)
if destination in self.trade_graph
else 0
)
features.extend([np.log1p(origin_volume), np.log1p(dest_volume)])
# HS code features (simplified)
hs_numeric = float(hs_code[:2]) if hs_code and hs_code[:2].isdigit() else 0
features.append(hs_numeric)
# Network centrality features
if origin in self.trade_graph:
origin_centrality = nx.degree_centrality(self.trade_graph)[origin]
else:
origin_centrality = 0
if destination in self.trade_graph:
dest_centrality = nx.degree_centrality(self.trade_graph)[destination]
else:
dest_centrality = 0
features.extend([origin_centrality, dest_centrality])
return features
def predict(self, policy_features) -> TradeFlowPrediction:
"""
Predict trade flow impacts from policy features
Args:
policy_features: Output from PolicyTriggerLayer
Returns:
TradeFlowPrediction: Predicted trade flow impacts
"""
if not self.is_fitted:
raise ValueError("Layer must be fitted before prediction")
# Extract relevant information from policy features
affected_countries = getattr(policy_features, "affected_countries", [])
hs_codes = getattr(policy_features, "hs_codes", [])
tariff_rates = getattr(policy_features, "tariff_rates", [])
country_tariff_map = getattr(policy_features, "country_tariff_map", {})
# Default to demo values if no policy features
if not affected_countries:
affected_countries = ["US", "CN"]
if not hs_codes:
hs_codes = ["85"]
if not tariff_rates:
tariff_rates = [0.25]
# Identify affected trade routes
affected_routes = self._identify_affected_routes(affected_countries, hs_codes)
# Predict flow changes for each route
flow_predictions = []
cost_predictions = []
for route in affected_routes:
features = self._extract_route_features(
route["origin"], route["destination"], route["hs_code"]
)
if len(features) > 0:
# Get country-specific tariff rate
country = (
route["origin"]
if route["origin"] in affected_countries
else route["destination"]
)
country_tariff_rate = country_tariff_map.get(country, 0.0)
# Scale features
features_scaled = self.feature_scaler.transform([features])
# Predict base impacts (use baseline models)
base_flow_change = self.flow_predictor.predict(features_scaled)[0]
base_cost_change = self.cost_predictor.predict(features_scaled)[0]
# Apply country-specific tariff rate directly
# Higher tariff = more flow disruption and cost increase
flow_change = base_flow_change * (
1 + country_tariff_rate * 2
) # Scale tariff impact
cost_change = country_tariff_rate * 0.8 # Direct tariff impact on costs
route["flow_change"] = flow_change
route["cost_change"] = cost_change
route["tariff_rate"] = country_tariff_rate
flow_predictions.append(flow_change)
cost_predictions.append(cost_change)
# Calculate aggregate impacts
supply_chain_disruption = self._calculate_supply_chain_disruption(
affected_routes
)
alternative_sources = self._find_alternative_sources(affected_routes)
reallocation_matrix = self._calculate_reallocation_matrix(affected_routes)
# Calculate cost increases and time delays
cost_increases = {
route["route_id"]: route.get("cost_change", 0) for route in affected_routes
}
time_delays = {
route["route_id"]: route.get("cost_change", 0) * 5
for route in affected_routes
} # Simplified
return TradeFlowPrediction(
affected_routes=affected_routes,
reallocation_matrix=reallocation_matrix,
supply_chain_disruption=supply_chain_disruption,
alternative_sources=alternative_sources,
cost_increases=cost_increases,
time_delays=time_delays,
)
def _identify_affected_routes(
self, countries: List[str], hs_codes: List[str]
) -> List[Dict[str, Any]]:
"""Identify trade routes affected by tariff changes"""
affected_routes = []
# Check all edges in trade graph
for origin, destination, edge_data in self.trade_graph.edges(data=True):
# Check if route involves affected countries
if origin in countries or destination in countries:
route = {
"route_id": f"{origin}-{destination}",
"origin": origin,
"destination": destination,
"hs_code": (
hs_codes[0] if hs_codes else "85"
), # Default to electronics
"baseline_volume": edge_data.get("trade_volume", 0),
"baseline_cost": edge_data.get("transport_cost", 0.05),
"baseline_time": edge_data.get("lead_time", 14),
}
affected_routes.append(route)
return affected_routes
def _calculate_supply_chain_disruption(
self, affected_routes: List[Dict[str, Any]]
) -> Dict[str, float]:
"""Calculate supply chain disruption scores by sector"""
disruption_scores = {}
# Group by HS code (sector)
sectors = {}
for route in affected_routes:
hs_code = route["hs_code"]
if hs_code not in sectors:
sectors[hs_code] = []
sectors[hs_code].append(route)
# Calculate disruption for each sector
for hs_code, routes in sectors.items():
total_impact = sum(abs(route.get("flow_change", 0)) for route in routes)
total_volume = sum(route["baseline_volume"] for route in routes)
if total_volume > 0:
disruption_score = min(total_impact / total_volume, 1.0)
else:
disruption_score = 0.0
disruption_scores[hs_code] = disruption_score
return disruption_scores
def _find_alternative_sources(
self, affected_routes: List[Dict[str, Any]]
) -> Dict[str, List[str]]:
"""Find alternative sourcing options for disrupted routes"""
alternatives = {}
for route in affected_routes:
destination = route["destination"]
origin = route["origin"]
# Find other countries that export to the same destination
alternative_origins = []
for node in self.trade_graph.nodes():
if (
node != origin
and self.trade_graph.has_edge(node, destination)
and node not in [origin]
):
alternative_origins.append(node)
alternatives[route["route_id"]] = alternative_origins[
:5
] # Top 5 alternatives
return alternatives
def _calculate_reallocation_matrix(
self, affected_routes: List[Dict[str, Any]]
) -> np.ndarray:
"""Calculate trade flow reallocation matrix"""
num_countries = len(self.country_to_index)
reallocation_matrix = np.zeros((num_countries, num_countries))
for route in affected_routes:
origin_idx = self.country_to_index.get(route["origin"])
dest_idx = self.country_to_index.get(route["destination"])
if origin_idx is not None and dest_idx is not None:
flow_change = route.get("flow_change", 0)
reallocation_matrix[origin_idx, dest_idx] = flow_change
return reallocation_matrix
def get_network_statistics(self) -> Dict[str, Any]:
"""Get trade network statistics"""
if not self.trade_graph:
return {}
return {
"num_countries": self.trade_graph.number_of_nodes(),
"num_routes": self.trade_graph.number_of_edges(),
"network_density": nx.density(self.trade_graph),
"average_clustering": nx.average_clustering(self.trade_graph),
"most_connected_countries": [
node
for node, degree in sorted(
list(self.trade_graph.degree()), key=lambda x: x[1], reverse=True
)[:5]
],
}
def fit(self, X, y=None):
"""Fit the trade flow model."""
pass
def predict(self, X):
"""Predict trade flow impacts."""
return None
|