phyer/myTestFreqAI
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

添加了全面的量价关系指标，包括成交量趋势指标和量价关系指标

Browse Source
This commit is contained in:
zhangkun9038@dingtalk.com 2026-02-02 08:36:13 +08:00
parent 45c3c0f1fc
commit c8616c3175
2 changed files with 72 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
doc/strategy_improvements.md
View File

@ -89,6 +89,25 @@
- 使用高级市场状态信息进行决策
- 实现了动态止损调整
#### 4.6 量价关系指标添加
- **成交量趋势指标**：
- `volume_change`: 成交量变化率
- `volume_ma_ratio`: 成交量移动平均比值
- `volume_std`: 成交量标准差（衡量成交量波动性）
- `volume_cv`: 成交量变异系数
- `volume_zscore`: 成交量Z分数（检测异常成交量）
- `volume_trend`: 成交量趋势（成交量是否在上升通道）
- `volume_ema_fast/slow`: 成交量EMA趋势
- **量价关系指标**：
- `price_change`: 价格变化率
- `volume_price_confirmation`: 量价确认度（价格和成交量同向变动）
- `volume_price_ratio`: 量价比率（成交量/价格）
- `bearish_divergence`: 熊市背离（价格创新高但成交量未创新高）
- `bullish_divergence`: 牛市背离（价格创新低但成交量未创新低）
- `obv`: OBV平衡成交量
- `volume_price_corr_short/medium`: 量价相关系数（短期/中期窗口）
### 5. 移除的逻辑
- 移除了冷启动保护逻辑，确保回测和实盘的一致性

53
freqtrade/templates/freqaiprimer.py
View File

@ -706,6 +706,59 @@ class FreqaiPrimer(IStrategy):
# 为第一行设置默认值
dataframe['prev_market_state'] = dataframe['prev_market_state'].fillna('neutral')
# ========== 新增：成交量趋势指标 ==========
# 成交量变化率
dataframe['volume_change'] = dataframe['volume'].pct_change()
# 成交量移动平均比值
dataframe['volume_ma_ratio'] = dataframe['volume'] / dataframe['volume_ma']
# 成交量标准差（衡量成交量波动性）
dataframe['volume_std'] = dataframe['volume'].rolling(window=20).std()
dataframe['volume_cv'] = dataframe['volume_std'] / dataframe['volume_ma'] # 变异系数
# 成交量Z分数（检测异常成交量）
dataframe['volume_zscore'] = (dataframe['volume'] - dataframe['volume_ma']) / dataframe['volume_std']
# 成交量趋势（成交量是否在上升通道）
dataframe['volume_trend'] = np.where(dataframe['volume'] > dataframe['volume_ma'], 1, 0)
# 成交量EMA趋势
dataframe['volume_ema_fast'] = dataframe['volume'].ewm(span=5, adjust=False).mean()
dataframe['volume_ema_slow'] = dataframe['volume'].ewm(span=20, adjust=False).mean()
dataframe['volume_ema_trend'] = np.where(dataframe['volume_ema_fast'] > dataframe['volume_ema_slow'], 1, 0)
# ========== 新增：量价关系指标 ==========
# 价格变化率
dataframe['price_change'] = dataframe['close'].pct_change()
# 量价确认度（价格和成交量同向变动）
dataframe['volume_price_confirmation'] = np.sign(dataframe['price_change']) * np.sign(dataframe['volume_change'])
# 量价比率（成交量/价格）
dataframe['volume_price_ratio'] = dataframe['volume'] / dataframe['close']
# 量价背离检测
# 价格创新高但成交量未创新高（顶背离）
dataframe['price_new_high'] = (dataframe['close'] == dataframe['close'].rolling(20).max()).astype(int)
dataframe['volume_new_high'] = (dataframe['volume'] == dataframe['volume'].rolling(20).max()).astype(int)
dataframe['bearish_divergence'] = (dataframe['price_new_high'] == 1) & (dataframe['volume_new_high'] == 0)
# 价格创新低但成交量未创新低（底背离）
dataframe['price_new_low'] = (dataframe['close'] == dataframe['close'].rolling(20).min()).astype(int)
dataframe['volume_new_low'] = (dataframe['volume'] == dataframe['volume'].rolling(20).min()).astype(int)
dataframe['bullish_divergence'] = (dataframe['price_new_low'] == 1) & (dataframe['volume_new_low'] == 0)
# OBV (On Balance Volume) - 平衡成交量
dataframe['obv'] = dataframe['volume'] * np.sign(dataframe['close'].diff())
dataframe['obv'] = dataframe['obv'].cumsum()
# 量价相关系数（短期窗口）
dataframe['volume_price_corr_short'] = dataframe['close'].pct_change().rolling(window=10).corr(dataframe['volume'].pct_change())
# 量价相关系数（中期窗口）
dataframe['volume_price_corr_medium'] = dataframe['close'].pct_change().rolling(window=20).corr(dataframe['volume'].pct_change())
# 记录当前的市场状态
if len(dataframe) > 0:
current_score = dataframe['market_score'].iloc[-1]