高频交易核心：限价订单簿(LOB)深度解析与量化策略

# 计算LOB深度的Python示例
import pandas as pd

def calculate_lob_depth(lob_df, levels=5):
    """计算LOB前N档深度"""
    bid_depth = lob_df['bid_size'].iloc[:levels].sum()
    ask_depth = lob_df['ask_size'].iloc[:levels].sum()
    return bid_depth, ask_depth

# 计算订单簿不平衡指标
def lob_imbalance(lob_df, levels=5):
    """LOB不平衡度：衡量买卖压力"""
    bid_vol = lob_df['bid_size'].iloc[:levels].sum()
    ask_vol = lob_df['ask_size'].iloc[:levels].sum()
    return (bid_vol - ask_vol) / (bid_vol + ask_vol)

# LOB不平衡策略示例
class OBImbalanceStrategy:
    def __init__(self, threshold=0.3, levels=5):
        self.threshold = threshold
        self.levels = levels
        
    def generate_signal(self, lob_snapshot):
        imbalance = lob_imbalance(lob_snapshot, self.levels)
        
        if imbalance > self.threshold:
            return "BUY"  # 买压强势，做多
        elif imbalance < -self.threshold:
            return "SELL"  # 卖压强势，做空
        else:
            return "HOLD"

# 简化版做市商策略
class SimpleMarketMaker:
    def __init__(self, spread_target=2):
        self.spread_target = spread_target  # 目标价差( ticks)
        
    def quote(self, lob):
        best_bid = lob['bid_price'].iloc[0]
        best_ask = lob['ask_price'].iloc[0]
        mid_price = (best_bid + best_ask) / 2
        
        # 在中间价附近挂双边单
        my_bid = mid_price - self.spread_target / 2
        my_ask = mid_price + self.spread_target / 2
        
        return {'bid': my_bid, 'ask': my_ask}

def detect_iceberg(lob_history, threshold=0.8):
    """检测可能的冰山订单"""
    signals = []
    
    for i in range(1, len(lob_history)):
        prev = lob_history[i-1]
        curr = lob_history[i]
        
        # 如果价格档位保持不变但数量频繁补充
        for level in range(5):
            if (prev['bid_price'].iloc[level] == curr['bid_price'].iloc[level] and
                curr['bid_size'].iloc[level] > prev['bid_size'].iloc[level] * threshold):
                signals.append(('ICEBERG_BUY', curr['bid_price'].iloc[level]))
                
    return signals