{"id":2196,"date":"2025-04-26T07:42:00","date_gmt":"2025-04-25T23:42:00","guid":{"rendered":"https:\/\/blog.laoyulaoyu.top\/?p=2196"},"modified":"2025-04-17T10:55:48","modified_gmt":"2025-04-17T02:55:48","slug":"%e8%ae%a9%e4%ba%a4%e6%98%93%e6%9c%ba%e5%99%a8%e4%ba%ba%e5%ad%a6%e4%bc%9a%e8%87%aa%e6%88%91%e8%bf%9b%e5%8c%96%ef%bc%9a%e6%88%91%e7%9a%84ai%e9%87%8f%e5%8c%96%e5%ae%9e%e6%88%98%e6%96%b0%e4%bd%93%e9%aa%8c","status":"publish","type":"post","link":"https:\/\/www.laoyulaoyu.com\/index.php\/2025\/04\/26\/%e8%ae%a9%e4%ba%a4%e6%98%93%e6%9c%ba%e5%99%a8%e4%ba%ba%e5%ad%a6%e4%bc%9a%e8%87%aa%e6%88%91%e8%bf%9b%e5%8c%96%ef%bc%9a%e6%88%91%e7%9a%84ai%e9%87%8f%e5%8c%96%e5%ae%9e%e6%88%98%e6%96%b0%e4%bd%93%e9%aa%8c\/","title":{"rendered":"\u8ba9\u4ea4\u6613\u673a\u5668\u4eba\u5b66\u4f1a\u81ea\u6211\u8fdb\u5316\uff1a\u6211\u7684AI\u91cf\u5316\u5b9e\u6218\u65b0\u4f53\u9a8c"},"content":{"rendered":"\n

\u4f5c\u8005\uff1a\u8001\u4f59\u635e\u9c7c<\/a><\/p>\n\n\n\n

\u539f\u521b\u4e0d\u6613\uff0c\u8f6c\u8f7d\u8bf7\u6807\u660e\u51fa\u5904\u53ca\u539f\u4f5c\u8005\u3002<\/mark><\/strong><\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n
\n
\u5199\u5728\u524d\u9762\u7684\u8bdd\uff1a<\/strong>\u6700\u8fd1\u670b\u53cb\u7ed9\u6211\u63a8\u8350\u4e86\u4e00\u4e2a\u53ebTensorTrade-NG\u7684\u667a\u80fd\u4ea4\u6613\u7cfb\u7edf\uff0c\u5b83\u662f\u4e2a\u4f1a\u7528AI\u2018\u81ea\u5b66\u2019\u7684\u4ea4\u6613\u673a\u5668\u4eba\uff01<\/mark>\u6211\u8bd5\u4e86\u8bd5\uff0c\u7ed3\u679c\u5f7b\u5e95\u4e0a\u5934\u2014\u2014\u6a21\u5757\u5316\u8bbe\u8ba1\u50cf\u642d\u4e50\u9ad8\uff0c\u96f6\u57fa\u7840\u4e5f\u80fd5\u5206\u949f\u8dd1\u901a\u4ee3\u7801\uff0c\u8fd8\u80fd\u7528\u5f3a\u5316\u5b66\u4e60\u8ba9\u673a\u5668\u4eba\u8d8a\u4e8f\u8d8a\u806a\u660e\u3002\u8fd9\u7bc7\u6587\u7ae0\uff0c\u6211\u4f1a\u624b\u628a\u624b\u6559\u4f60\u5982\u4f55\u7528\u5b83\u907f\u5f00\u97ed\u83dc\u547d\u8fd0\uff0c\u9644\u4ee3\u7801\u3001\u907f\u5751\u6307\u5357\u548c\u5b9e\u6218\u7ecf\u9a8c\uff01<\/pre>\n<\/blockquote>\n\n\n\n
\u4e00\u3001\u4f20\u7edf\u91cf\u5316\u4ea4\u6613\u673a\u5668\u4eba VS AI\u673a\u5668\u4eba\uff1a\u8c01\u66f4\u6297\u63cd\uff1f<\/strong><\/strong><\/h2>\n\n\n\n
\u4f20\u7edf\u91cf\u5316\u4ea4\u6613\u7cfb\u7edf\u5c31\u50cf\u53ea\u4f1a\u80cc\u4e66\u7684\u5b66\u9738\uff0c\u9047\u5230\u65b0\u9898\u5c31\u50bb\u773c\u3002\u6bd4\u5982\uff0c\u8bbe\u5b9a\u201c\u4ef7\u683c\u6da83%\u5c31\u4e70\u201d\uff0c\u4f46\u5e02\u573a\u4e00\u6ce2\u52a8\uff0c\u5b83\u5c31\u53ef\u80fd\u4e8f\u5f97\u627e\u4e0d\u7740\u5317\u3002<\/p>\n\n\n\n
\u60f3\u8c61\u4e00\u4e0b\uff0c\u5982\u679c\u4f60\u7684\u91cf\u5316\u673a\u5668\u4eba\u80fd\u50cf\u4eba\u4e00\u6837\uff0c\u4ece\u5931\u8d25\u4e2d\u5b66\u4e60\uff0c\u8d8a\u6765\u8d8a\u806a\u660e\u5462\uff1f<\/p>\n\n\n\n
\"\"<\/figure>\n\n\n\n
\u800cTensorTrade-NG\u673a\u5668\u4eba\u5c31\u662f\u201c\u8001\u53f8\u673a\u201d\u2014\u2014\u901a\u8fc7\u5f3a\u5316\u5b66\u4e60\uff08AI\u8bd5\u9519\u6cd5\uff09<\/strong>\uff0c\u8ba9\u4ea4\u6613\u673a\u5668\u4eba\u901a\u8fc7\u4e0d\u65ad\u8bd5\u9519\u6765\u6539\u5584\u81ea\u5df1\u7684\u4ea4\u6613\u7b56\u7565\u3002<\/p>\n\n\n\n
\u5b83\u8fd8\u8ddf Stable-Baselines3<\/a> \u8fd9\u6837\u7684\u5f3a\u5316\u5b66\u4e60\u5e93\u65e0\u7f1d\u5bf9\u63a5\uff0c\u8ba9\u4f60\u8f7b\u677e\u7528\u4e0a\u9876\u5c16\u7b97\u6cd5\u3002\u5982\u679c\u6570\u636e\u91cf\u5927\uff0c\u5b83\u8fd8\u80fd\u901a\u8fc7 Ray \u5206\u5e03\u5f0f\u8ba1\u7b97\uff0c\u8dd1\u5f97\u98de\u5feb\u3002<\/p>\n\n\n\n
\u6838\u5fc3\u4f18\u52bf\u5bf9\u6bd4\u8868\uff1a\u4e3a\u4ec0\u4e48\u5b83\u6bd4\u4f20\u7edf\u91cf\u5316\u5de5\u5177\u66f4\u9999\uff1f<\/strong><\/p>\n\n\n\n
\u529f\u80fd<\/strong><\/th>\u4f20\u7edf\u6846\u67b6<\/strong><\/th>TensorTrade-NG<\/strong><\/th><\/tr><\/thead>
\u5b66\u4e60\u80fd\u529b<\/strong><\/td>\u56fa\u5b9a\u89c4\u5219\uff0c\u5e02\u573a\u4e00\u53d8\u5c31\u5e9f<\/td>AI\u81ea\u5b66\uff0c\u8d8a\u4ea4\u6613\u8d8a\u806a\u660e<\/td><\/tr>
\u5f00\u53d1\u901f\u5ea6<\/strong><\/td>\u4ece\u96f6\u9020\u8f6e\u5b50\uff0c\u8017\u65f6\u6570\u5468<\/td>\u6a21\u5757\u5316\u62fc\u88c5\uff0c5\u5206\u949f\u8dd1\u901a\u4ee3\u7801<\/td><\/tr>
\u6269\u5c55\u6027<\/strong><\/td>\u4ec5\u652f\u6301\u7b80\u5355\u7b56\u7565<\/td>\u80a1\u7968\u3001\u52a0\u5bc6\u8d27\u5e01\u3001\u671f\u8d27\u5168\u652f\u6301<\/td><\/tr>
\u98ce\u9669\u63a7\u5236<\/strong><\/td>\u624b\u52a8\u8bbe\u7f6e\u6b62\u635f<\/td>\u81ea\u52a8\u5b66\u4e60\u89c4\u907f\u9ad8\u98ce\u9669\u64cd\u4f5c<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
\u4e8c\u3001TensorTrade-NG \u662f\u4ec0\u4e48\uff1f<\/strong><\/h2>\n\n\n\n
TensorTrade-NG \u662f\u4e00\u4e2a\u5f00\u6e90\u7684 Python \u6846\u67b6\uff0c\u4e13\u95e8\u7528\u6765\u6253\u9020\u3001\u8bad\u7ec3\u548c\u6d4b\u8bd5\u4ea4\u6613\u673a\u5668\u4eba\u3002\u8fd9\u4e2a\u673a\u5668\u4eba\u4e0d\u9760\u6b7b\u677f\u7684\u89c4\u5219\uff0c\u800c\u662f\u901a\u8fc7\u5f3a\u5316\u5b66\u4e60\uff08RL\uff09\u81ea\u5df1\u6478\u7d22\u600e\u4e48\u4ea4\u6613\u3002<\/p>\n\n\n\n
\"\"<\/figure>\n\n\n\n
TensorTrade-NG\u7684\u76ee\u6807\u5f88\u660e\u786e\u2014\u2014\u5e2e\u52a9\u5927\u5bb6\u642d\u5efa\u4e00\u4e2a\u5177\u6709\u5b66\u4e60\u80fd\u529b\u7684\u4ea4\u6613\u7cfb\u7edf\u3002<\/p>\n\n\n\n
\"\"<\/figure>\n\n\n\n
\u6838\u5fc3\u529f\u80fd<\/strong>\uff1a<\/p>\n\n\n\n
\"\"<\/figure>\n\n\n\n
\u8fd9\u4e2a\u5de5\u5177\u65e2\u652f\u6301\u6a21\u62df\u73af\u5883\uff0c\u4e5f\u80fd\u63a5\u5165\u771f\u5b9e\u5e02\u573a\u6570\u636e\u3002\u5b83\u80fd\u5904\u7406\u80a1\u7968\u3001\u52a0\u5bc6\u8d27\u5e01\u7b49\u5404\u79cd\u8d44\u4ea7\u3002\u6574\u4e2a\u67b6\u6784\u7075\u6d3b\u5f97\u50cf\u642d\u4e50\u9ad8\u79ef\u6728\uff0c\u6bcf\u4e00\u5757\u6a21\u5757\u90fd\u53ef\u4ee5\u81ea\u7531\u66ff\u6362\uff0c\u6ee1\u8db3\u4ece\u5165\u95e8\u5230\u8fdb\u9636\u5404\u79cd\u9700\u6c42\u3002<\/p>\n\n\n\n
\u76ee\u524d\u5b83\u8fd8\u5728 beta \u9636\u6bb5\uff0c\u53ef\u80fd\u7528\u8d77\u6765\u4f1a\u6709\u70b9\u5c0f\u95ee\u9898\u9700\u8981\u8c03\u6574\uff0c\u4f46\u6f5c\u529b\u5de8\u5927\u3002\u5b83\u590d\u7528\u4e86 NumPy \u548c Pandas \u8fd9\u4e9b Python \u795e\u5668\uff0c\u5e2e\u4f60\u7701\u53bb\u5f88\u591a\u91cd\u590d\u5de5\u4f5c\u3002<\/p>\n\n\n\n
\u4f60\u53ef\u4ee5\u5feb\u901f\u8bd5\u9519\uff0c\u518d\u4e5f\u4e0d\u7528\u82b1\u51e0\u5468\u65f6\u95f4\u4ece\u5934\u5199\u4ee3\u7801\u3002<\/p>\n\n\n\n
\u4e09\u3001\u624b\u628a\u624b\u6559\u4f60\u642d\u5efa\u7b2c\u4e00\u4e2aAI\u4ea4\u6613\u673a\u5668\u4eba<\/strong><\/h2>\n\n\n\n

3.1  \u73af\u5883\u914d\u7f6e<\/strong><\/h3>\n\n\n\n
\u5b89\u88c5\u53ea\u9700\u4e00\u884c\u4ee3\u7801\uff0c\u8fdePython\u5c0f\u767d\u90fd\u80fd\u79d2\u4e0a\u624b\uff1a<\/p>\n\n\n\n
pip install tensortrade-ng  # \u7a33\u5982\u8001\u72d7\u7684\u7a33\u5b9a\u7248  \n# \u6216\u6311\u6218\u6781\u9650\uff01\u88c5\u6700\u65b0\u5f00\u53d1\u7248\uff1a  \npip install git+https:\/\/github.com\/erhardtconsulting\/tensortrade-ng.git  <\/code><\/pre>\n\n\n\n
\u907f\u5751\u63d0\u793a<\/strong>\uff1a\u8bb0\u5f97\u5148\u88c5git-lfs<\/code>\uff0c\u5426\u5219\u4f1a\u5361\u5728\u6570\u636e\u4e0b\u8f7d\uff01Mac\u7528brew install git-lfs<\/code>\uff0cWin\u73af\u5883\u53bb\u5b98\u7f51\u4e0b\u3002<\/p>\n\n\n\n
\u88c5\u597d\u540e\uff0c\u6d4b\u8bd5\u4e00\u4e0b\uff1a<\/p>\n\n\n\n
hatch run test:run<\/code><\/code><\/pre>\n\n\n\n
\u60f3\u770b\u6587\u6863\uff1f\u672c\u5730\u8dd1\uff1a<\/p>\n\n\n\n
hatch run docs:serve<\/code><\/code><\/pre>\n\n\n\n
\u6587\u6863\u5199\u5f97\u6e05\u695a\uff0c\u6551\u4e86\u6211\u597d\u51e0\u6b21\u3002\u9700\u8981 pandoc \u7684\u8bdd\uff0c\u522b\u7528 pip \u88c5\uff0c\u7528 brew \u6216 apt \u641e\u5b9a\u3002<\/p>\n\n\n\n
3.2 \u4e94\u5206\u949f\u4ee3\u7801\u5b9e\u6218\uff1a\u8ba9AI\u4e70\u5356\u865a\u62df\u80a1\u7968<\/strong><\/h3>\n\n\n\n
\u4e0b\u9762\u8fd9\u6bb5\u4ee3\u7801\u7b80\u5355\u7c97\u66b4\uff0c\u4f46\u4f60\u80fd\u770b\u5230\u73af\u5883\u5df2\u7ecf\u642d\u597d\uff0c\u673a\u5668\u4eba\u5f00\u59cb\u4ea4\u6613\u4e86\u3002<\/p>\n\n\n\n
import tensortrade_ng as tt  \nfrom tensortrade_ng.env import TradingEnv  \nfrom tensortrade_ng.exchanges import SimulatedExchange  \nimport pandas as pd  \n\n# \u6a21\u62df\u80a1\u4ef7\u6570\u636e\uff085\u5929\uff09  \ndata = pd.DataFrame({\"price\": [100, 102, 101, 105, 103]})  \nexchange = SimulatedExchange(data)  # \u865a\u62df\u4ea4\u6613\u6240  \n\n# \u8bbe\u5b9a\u4e70\u5356\u89c4\u5219\u548c\u5956\u52b1  \nactions = tt.actions.SimpleOrders()  # \u652f\u6301\u4e70\u3001\u5356\u3001\u8eba\u5e73  \nrewards = tt.rewards.SimpleProfit()  # \u8d5a\u94b1\u624d\u662f\u786c\u9053\u7406  \n\n# \u7ec4\u88c5\u4ea4\u6613\u73af\u5883  \nenv = TradingEnv(exchange=exchange, action_scheme=actions, reward_scheme=rewards)  \n\n# \u8ba9AI\u968f\u673a\u8bd5\u6c34\uff01  \nobs = env.reset()  \nfor _ in range(5):  \n    action = env.action_space.sample()  # \u968f\u673a\u9009\u52a8\u4f5c\uff080=\u8eba\u5e73\uff0c1=\u4e70\uff0c2=\u5356\uff09  \n    obs, reward, done, info = env.step(action)  \n    print(f\"\u64cd\u4f5c\uff1a{action}\uff0c\u8d5a\u4e86\uff1a{reward}\u5143\uff01\")  <\/code><\/pre>\n\n\n\n
\u8fd0\u884c\u7ed3\u679c\u793a\u4f8b<\/strong>\uff1a<\/p>\n\n\n\n
\u64cd\u4f5c\uff1a1\uff0c\u8d5a\u4e86\uff1a2\u5143\uff01  \n\u64cd\u4f5c\uff1a2\uff0c\u8d5a\u4e86\uff1a-1\u5143\uff01  \n\u64cd\u4f5c\uff1a0\uff0c\u8d5a\u4e86\uff1a0\u5143\uff01\uff08\u8eba\u5e73\u53ef\u803b\u4f46\u6709\u7528\uff09  <\/code><\/pre>\n\n\n\n
3. 3 \u8fdb\u9636\u73a9\u6cd5\uff1a\u63a5\u5165\u771f\u5b9e\u5e02\u573a\u6570\u636e<\/strong><\/h3>\n\n\n\n
\u5728AI\u91cf\u5316\u5b9e\u6218\u4e2d\uff0c\u5982\u679c\u4f60\u8fd8\u505c\u7559\u5728 yfinance<\/code> \u4e0b\u8f7d\u5386\u53f2\u6570\u636e\u7684\u9636\u6bb5\uff0c\u90a3\u4f60\u5c31\u9519\u8fc7\u4e86\u771f\u6b63\u7684\u4ea4\u6613\u673a\u4f1a\u3002<\/p>\n\n\n\n
QMT\uff08\u5238\u5546\u91cf\u5316\u4ea4\u6613\u7ec8\u7aef\uff09+ PTrade\uff08\u4e8b\u4ef6\u9a71\u52a8\u7b56\u7565\u5f15\u64ce\uff09\u7ed9\u6211\u4eec\u63d0\u4f9b\u4e86\u66f4\u9ad8\u7ef4\u5ea6\u7684\u771f\u5b9e\u6570\u636e\u4e0e\u7b56\u7565\u6267\u884c\u80fd\u529b<\/strong>\u3002\u53ea\u9700\u5f00\u901aQMT\u91cf\u5316\u6743\u9650\uff0c\u6211\u4eec\u5c31\u80fd\u7528 Python \u76f4\u63a5\u8bfb\u53d6\u8d26\u6237\u6570\u636e\u3001\u5b9e\u65f6\u884c\u60c5\uff0c\u751a\u81f3\u6302\u5355\u6210\u4ea4\u3002<\/p>\n\n\n\n
\u4ee5\u4e0b\u662f\u4e00\u4e2a\u901a\u8fc7 PTrade \u5e73\u53f0\u83b7\u53d6\u884c\u60c5\u6570\u636e\u7684\u9ad8\u6548\u65b9\u5f0f\uff1a<\/p>\n\n\n\n
# \u83b7\u53d6\u5386\u53f2\u884c\u60c5\u6570\u636e\u793a\u4f8b\uff08\u65e0\u9700\u518d\u7528 yfinance\uff09\ndf = get_history(10, '1d', 'close', security='600519.SS', fq=None, include=False)\nprint(df)<\/code><\/pre>\n\n\n\n
\u56db\u3001\u6211\u7684\u5b9e\u6218\u4f53\u9a8c<\/strong><\/h2>\n\n\n\n
\u521d\u5c1d TensorTrade-NG \u7684\u65f6\u5019\uff0c\u6211\u8fd8\u6c89\u8ff7\u5728\u201cAI \u4f1a\u8d8a\u6765\u8d8a\u806a\u660e\u201d\u7684\u5e7b\u60f3\u4e2d\u3002\u867d\u7136\u8fc7\u7a0b\u5076\u5c14\u9047\u5230\u5c0f\u95ee\u9898\uff0c\u6574\u4f53\u4f53\u9a8c\u8fd8\u7b97\u975e\u5e38\u987a\u7545\u3002<\/p>\n\n\n\n
\u4f46\u5f53\u6211\u771f\u6b63\u63a5\u5165 QMT+PTrade \u7684\u7ec4\u5408\u540e\uff0c\u6211\u624d\u610f\u8bc6\u5230 \u2014\u2014 \u771f\u6b63\u806a\u660e\u7684\uff0c\u4e0d\u53ea\u662f AI\uff0c\u800c\u662f\u7528\u5bf9\u4e86\u5de5\u5177\u7684\u4eba\u7c7b\u81ea\u5df1<\/strong>\u3002<\/p>\n\n\n\n
\"\"<\/figure>\n\n\n\n
\u6211\u7684\u7b56\u7565\u843d\u5730\u6d41\u7a0b\u5982\u4e0b\uff1a<\/p>\n\n\n\n
\u7b56\u7565\u5f00\u53d1\uff1a<\/strong> \u4f7f\u7528 PTrade \u7684\u4e8b\u4ef6\u9a71\u52a8\u6846\u67b6\uff0c\u5b9a\u4e49\u597d initialize<\/code>\u3001handle_data<\/code> \u7b49\u51fd\u6570\uff0c\u901a\u8fc7 get_history<\/code> \u6293\u53d6\u771f\u5b9e\u884c\u60c5\uff0c\u5b9e\u65f6\u6267\u884c\u4e70\u5356\u51b3\u7b56\u3002<\/p>\n\n\n\n
\u672c\u5730\u8fde\u63a5 QMT\uff1a<\/strong> \u901a\u8fc7 Python \u811a\u672c\u8fde\u63a5 miniqmt<\/code> \u5ba2\u6237\u7aef\uff0c\u5b9e\u73b0\u8d26\u6237\u767b\u5f55\u3001\u8d44\u4ea7\u67e5\u8be2\u3001\u6301\u4ed3\u76d1\u63a7\u3002<\/p>\n\n\n\n
from xtquant.xttrader import XtQuantTrader\ntrader = XtQuantTrader(r'D:\\QMT\u8def\u5f84', session_id=123456)\ntrader.start()\ntrader.connect()<\/code><\/pre>\n\n\n\n
\u8d26\u6237\u7ba1\u7406\uff1a<\/strong> \u901a\u8fc7\u8c03\u7528 query_stock_asset()<\/code>\u3001query_stock_positions()<\/code> \u7b49\u63a5\u53e3\uff0c\u968f\u65f6\u638c\u63e1\u8d44\u91d1\u4e0e\u4ed3\u4f4d\u3002<\/p>\n\n\n\n
\"\"<\/figure>\n\n\n\n
\u5b9e\u76d8\u4ea4\u6613\uff1a<\/strong> \u56de\u6d4b\u901a\u8fc7\u540e\uff0c\u5207\u6362\u5230\u5b9e\u76d8\u63a5\u53e3\uff0c\u7b56\u7565\u4ee3\u7801\u4e0d\u53d8\uff0c\u53ea\u9700\u66f4\u6362 API \u8fde\u63a5\u8bbe\u7f6e\u5373\u53ef\u5f00\u59cb\u5b9e\u76d8\u64cd\u4f5c\u3002<\/p>\n\n\n\n
\"\"<\/figure>\n\n\n\n
\u5b89\u5168\u6027\u4e0e\u7a33\u5b9a\u6027\uff1a<\/strong><\/p>\n\n\n\n