Spaces:
Running
Running
File size: 10,246 Bytes
c095e08 |
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 |
"""
NBA ML Prediction System - MVP Predictor
=========================================
Model to predict MVP based on player performance, team success,
and historical MVP similarity.
"""
import numpy as np
import pandas as pd
from pathlib import Path
from typing import Dict, List, Optional
from sklearn.metrics.pairwise import cosine_similarity
from sklearn.preprocessing import StandardScaler
import xgboost as xgb
import joblib
import logging
from src.config import MODELS_DIR, RAW_DATA_DIR
logger = logging.getLogger(__name__)
# =============================================================================
# HISTORICAL MVP PROFILES
# =============================================================================
# Historical MVP seasons (approximate stats for similarity comparison)
HISTORICAL_MVP_PROFILES = {
"2023-24": {"player": "Nikola Jokic", "ppg": 26.4, "rpg": 12.4, "apg": 9.0, "ws": 17.8, "team_wins": 57},
"2022-23": {"player": "Joel Embiid", "ppg": 33.1, "rpg": 10.2, "apg": 4.2, "ws": 14.3, "team_wins": 54},
"2021-22": {"player": "Nikola Jokic", "ppg": 27.1, "rpg": 13.8, "apg": 7.9, "ws": 15.2, "team_wins": 48},
"2020-21": {"player": "Nikola Jokic", "ppg": 26.4, "rpg": 10.8, "apg": 8.3, "ws": 15.6, "team_wins": 47},
"2019-20": {"player": "Giannis Antetokounmpo", "ppg": 29.5, "rpg": 13.6, "apg": 5.6, "ws": 14.4, "team_wins": 56},
"2018-19": {"player": "Giannis Antetokounmpo", "ppg": 27.7, "rpg": 12.5, "apg": 5.9, "ws": 14.4, "team_wins": 60},
"2017-18": {"player": "James Harden", "ppg": 30.4, "rpg": 5.4, "apg": 8.8, "ws": 15.4, "team_wins": 65},
"2016-17": {"player": "Russell Westbrook", "ppg": 31.6, "rpg": 10.7, "apg": 10.4, "ws": 13.1, "team_wins": 47},
"2015-16": {"player": "Stephen Curry", "ppg": 30.1, "rpg": 5.4, "apg": 6.7, "ws": 17.9, "team_wins": 73},
}
# =============================================================================
# MVP PREDICTOR
# =============================================================================
class MVPPredictor:
"""
Predicts MVP vote share using gradient boosting with narrative features.
"""
def __init__(self):
self.model = None
self.scaler = StandardScaler()
self.feature_columns = None
self.trained = False
def calculate_mvp_similarity(self, player_stats: Dict) -> float:
"""
Calculate cosine similarity to historical MVP profiles.
Captures voter psychology by finding players who "look like" past MVPs.
"""
# Create feature vector for player
player_vector = np.array([
player_stats.get("ppg", 0),
player_stats.get("rpg", 0),
player_stats.get("apg", 0),
player_stats.get("ws", 0),
player_stats.get("team_wins", 0) / 82 # Normalize to 0-1
]).reshape(1, -1)
# Create matrix of historical MVP profiles
mvp_vectors = []
for season, profile in HISTORICAL_MVP_PROFILES.items():
mvp_vectors.append([
profile["ppg"],
profile["rpg"],
profile["apg"],
profile["ws"],
profile["team_wins"] / 82
])
mvp_matrix = np.array(mvp_vectors)
# Normalize
if len(mvp_matrix) > 0:
mvp_matrix_normalized = self.scaler.fit_transform(mvp_matrix)
player_normalized = self.scaler.transform(player_vector)
# Calculate similarity to each MVP season
similarities = cosine_similarity(player_normalized, mvp_matrix_normalized)[0]
# Return max similarity (closest to any MVP)
return float(np.max(similarities))
return 0.0
def calculate_narrative_features(self, player_stats: Dict,
prev_season_stats: Optional[Dict] = None) -> Dict:
"""
Calculate narrative momentum features that voters care about.
"""
features = {}
# Stat improvement year-over-year
if prev_season_stats:
features["ppg_improvement"] = player_stats.get("ppg", 0) - prev_season_stats.get("ppg", 0)
features["rpg_improvement"] = player_stats.get("rpg", 0) - prev_season_stats.get("rpg", 0)
features["apg_improvement"] = player_stats.get("apg", 0) - prev_season_stats.get("apg", 0)
else:
features["ppg_improvement"] = 0
features["rpg_improvement"] = 0
features["apg_improvement"] = 0
# Team success
features["team_wins"] = player_stats.get("team_wins", 0)
features["team_win_pct"] = player_stats.get("team_wins", 41) / 82
# Games played (durability matters)
features["games_played"] = player_stats.get("gp", 0)
features["games_played_pct"] = player_stats.get("gp", 0) / 82
return features
def prepare_features(self, player_df: pd.DataFrame) -> pd.DataFrame:
"""Prepare all features for MVP prediction."""
features = player_df.copy()
# Calculate MVP similarity for each player
features["mvp_similarity"] = features.apply(
lambda row: self.calculate_mvp_similarity({
"ppg": row.get("PTS", 0),
"rpg": row.get("REB", 0),
"apg": row.get("AST", 0),
"ws": row.get("WS", 10), # Default if not available
"team_wins": row.get("TEAM_WINS", 41)
}), axis=1
)
return features
def train(self, X: np.ndarray, y: np.ndarray, feature_columns: List[str]):
"""Train the MVP prediction model."""
self.feature_columns = feature_columns
self.model = xgb.XGBRegressor(
n_estimators=200,
max_depth=5,
learning_rate=0.1,
random_state=42
)
self.model.fit(X, y)
self.trained = True
logger.info("MVP model trained")
def predict_vote_share(self, X: np.ndarray) -> np.ndarray:
"""Predict MVP vote share (0-1 scale)."""
if not self.trained:
raise ValueError("Model not trained")
return self.model.predict(X)
def rank_candidates(self, player_df: pd.DataFrame, top_n: int = 10) -> pd.DataFrame:
"""
Rank MVP candidates and return top N.
Uses real stats-based scoring formula.
"""
df = player_df.copy()
# MVP score based on stats available from NBA API
# Weighted formula considering:
# - Scoring (30%): Points per game
# - Playmaking (20%): Assists per game
# - Rebounding (15%): Rebounds per game
# - Defense (10%): Steals + Blocks
# - Efficiency (10%): Plus/Minus and FG%
# - Team Success (15%): Team win percentage
pts = df.get("PTS", pd.Series([0]*len(df))).fillna(0)
ast = df.get("AST", pd.Series([0]*len(df))).fillna(0)
reb = df.get("REB", pd.Series([0]*len(df))).fillna(0)
stl = df.get("STL", pd.Series([0]*len(df))).fillna(0)
blk = df.get("BLK", pd.Series([0]*len(df))).fillna(0)
plus_minus = df.get("PLUS_MINUS", pd.Series([0]*len(df))).fillna(0)
fg_pct = df.get("FG_PCT", pd.Series([0.45]*len(df))).fillna(0.45)
team_win_pct = df.get("TEAM_WIN_PCT", pd.Series([0.5]*len(df))).fillna(0.5)
df["mvp_score"] = (
pts * 1.0 + # Points (raw weight)
ast * 2.0 + # Assists (weighted more for playmaking)
reb * 1.0 + # Rebounds
(stl + blk) * 1.5 + # Defense
plus_minus * 0.3 + # Impact metric
fg_pct * 20 + # Efficiency bonus
team_win_pct * 30 # Team success (big factor for MVP)
)
# Add MVP similarity if we can calculate it
if "mvp_similarity" not in df.columns:
df = self.prepare_features(df)
if "mvp_similarity" in df.columns:
df["mvp_score"] = df["mvp_score"] + df["mvp_similarity"].fillna(0) * 10
# Sort and return top candidates
df = df.sort_values("mvp_score", ascending=False)
# Ensure columns exist for return
if "mvp_similarity" not in df.columns:
df["mvp_similarity"] = 0.0
return df.head(top_n)[["PLAYER_NAME", "PTS", "REB", "AST", "mvp_score", "mvp_similarity"]]
def save(self, path: Path = None):
"""Save model to disk."""
if path is None:
path = MODELS_DIR / "mvp_predictor.joblib"
joblib.dump({
"model": self.model,
"scaler": self.scaler,
"feature_columns": self.feature_columns,
"trained": self.trained
}, path)
logger.info(f"Saved MVP model to {path}")
def load(self, path: Path = None):
"""Load model from disk."""
if path is None:
path = MODELS_DIR / "mvp_predictor.joblib"
data = joblib.load(path)
self.model = data["model"]
self.scaler = data["scaler"]
self.feature_columns = data["feature_columns"]
self.trained = data["trained"]
# =============================================================================
# CLI INTERFACE
# =============================================================================
if __name__ == "__main__":
print("Testing MVP Similarity Calculator...")
predictor = MVPPredictor()
# Test with a hypothetical MVP-caliber season
test_stats = {
"ppg": 28.5,
"rpg": 12.0,
"apg": 8.5,
"ws": 15.0,
"team_wins": 55
}
similarity = predictor.calculate_mvp_similarity(test_stats)
print(f"MVP Similarity Score: {similarity:.3f}")
# Test narrative features
prev_stats = {"ppg": 25.0, "rpg": 10.0, "apg": 7.0}
narrative = predictor.calculate_narrative_features(test_stats, prev_stats)
print(f"Narrative Features: {narrative}")
|