71 lines
2.7 KiB
Python
71 lines
2.7 KiB
Python
# ============================================================
|
||
# 截面策略指标示例(研究参考版)
|
||
# Cross-Sectional Strategy Indicator Example (Research Reference)
|
||
# Momentum + RSI Composite Score
|
||
# ============================================================
|
||
#
|
||
# 使用方法:
|
||
# 1. 作为截面研究思路示例阅读
|
||
# 2. 用于理解“对多个标的打分再排序”的基本写法
|
||
#
|
||
# 当前限制:
|
||
# - 策略快照回测(cross_sectional)暂不支持;实盘指标截面策略可在「交易助手」创建
|
||
# - 创建时选择「截面策略」并插入模板,或在本指标中实现 scores 字典
|
||
#
|
||
# 评分逻辑:
|
||
# - 动量因子(20周期):价格变化率,越高越好
|
||
# - RSI 指标(14周期):反转 RSI 值,越低越好(100 - RSI)
|
||
# - 综合评分:70% 动量 + 30% RSI 反转值
|
||
#
|
||
# ============================================================
|
||
|
||
# 截面策略指标
|
||
# 输入: data = {symbol1: df1, symbol2: df2, ...}
|
||
# 输出: scores = {symbol1: score1, symbol2: score2, ...}
|
||
|
||
scores = {}
|
||
|
||
# 遍历全部标的
|
||
for symbol, df in data.items():
|
||
# 确保数据长度足够
|
||
if len(df) < 20:
|
||
scores[symbol] = 0
|
||
continue
|
||
|
||
# === 1. 计算动量因子 (20周期) ===
|
||
# 动量 = (当前价格 / 20周期前价格 - 1) * 100
|
||
momentum = (df['close'].iloc[-1] / df['close'].iloc[-20] - 1) * 100
|
||
|
||
# === 2. 计算RSI指标 (14周期) ===
|
||
def calculate_rsi(prices, period=14):
|
||
"""计算 RSI 指标"""
|
||
delta = prices.diff()
|
||
gain = (delta.where(delta > 0, 0)).rolling(window=period).mean()
|
||
loss = (-delta.where(delta < 0, 0)).rolling(window=period).mean()
|
||
rs = gain / loss
|
||
rsi = 100 - (100 / (1 + rs))
|
||
return rsi.iloc[-1]
|
||
|
||
rsi_value = calculate_rsi(df['close'], 14)
|
||
|
||
# === 3. 综合评分 ===
|
||
# 动量越高 = 评分越高
|
||
# RSI越低(超卖)= 评分越高(100 - RSI)
|
||
# 权重: 70% 动量 + 30% RSI反转值
|
||
momentum_score = momentum
|
||
rsi_score = 100 - rsi_value # 反转 RSI(RSI 越低,评分越高)
|
||
|
||
composite_score = momentum_score * 0.7 + rsi_score * 0.3
|
||
|
||
scores[symbol] = composite_score
|
||
|
||
# === 可选:手动指定排序 ===
|
||
# 如果不提供,系统会根据 scores 自动排序
|
||
# rankings = sorted(scores.keys(), key=lambda x: scores[x], reverse=True)
|
||
|
||
# === 研究语义说明 ===
|
||
# 1. 根据评分对所有标的进行排序(从高到低)
|
||
# 2. 选择排名靠前的 N 个标的做多(基于 portfolio_size * long_ratio)
|
||
# 3. 选择排名靠后的 N 个标的做空(基于 portfolio_size * (1 - long_ratio))
|
||
# 4. 在未来平台链路完善后,可由系统统一生成买入 / 卖出 / 平仓动作
|