---
excalidraw-plugin: parsed
tags: [excalidraw, quants, regression, alpha-model, quant-architecture]
---
==⚠  Switch to EXCALIDRAW VIEW in the MORE OPTIONS menu of this document. ⚠==

# Excalidraw Data

## Text Elements
一笔量化交易是怎样生产出来的 ^title

回归和损失函数只位于 Alpha 模型内部；赚钱是系统级结果，不是单一模型按钮。 ^subtitle

① 数据层 ^dataTitle

Point-in-time 行情
财务 / 事件 / 另类数据
复权、对齐、去未来函数
标签与训练/验证/测试切分 ^dataText

② Alpha 模型 ^alphaTitle

特征 → 预期收益 / 排名
OLS、树模型、深度学习
损失函数在这里训练模型 ^alphaText

你当前在这里：OLS → 损失 ^currentText

③ 风险模型 ^riskTitle

暴露 / 协方差 / 特异风险
行业、个股、杠杆与回撤约束 ^riskText

④ 成本模型 ^costTitle

佣金 / 印花税 / 价差
冲击、滑点、容量与换手 ^costText

一笔量化交易是怎样生产出来的 ^title

① 数据层 ^utGzA754

② Alpha 模型 ^xXUylBPu

特征 → 预期收益 / 排名
OLS、树模型、深度学习
损失函数在这里训练模型 ^ATcM6KYG

你当前在这里：OLS → 损失 ^RgBMmEgg

③ 风险模型 ^RYcyQs0x

④ 成本模型 ^TeVlProy

⑤ 组合构建 ^52ubecvM

最大化预期收益
− 风险惩罚
− 交易成本
并满足仓位、行业、流动性约束
输出：目标权重 ^JxtaptoE

⑥ 执行模型 ^QzmEx4JX

目标权重 → 订单
拆单、限价、时机、成交
输出：真实持仓与现金 ^CZ33e3kn

⑦ 归因、监控与研究反馈 ^ocZx0fRj

成本前后收益 / 风险暴露 / 失效诊断
反馈到数据、Alpha、风险与成本 ^RRfdntwg

模型损失决定“怎样拟合标签”；组合目标决定“怎样取舍收益、风险和成本”；执行目标决定“怎样更便宜地成交”。三者不是同一个目标。 ^legend

%%
## Drawing
```compressed-json
N4KAkARALgngDgUwgLgAQQQDwMYEMA2AlgCYBOuA7hADTgQBuCpAzoQPYB2KqATLZMzYBXUtiRoIACyhQ4zZAHoFAc0JRJQgEYA6bGwC2CgF7N6hbEcK4OCtptbErHALRY8RMpWdx8Q1TdIEfARcZgRmBShcZQUebR4AVniaOiCEfQQOKGZuAG1wMFAwYogSbghiXCiAITZMFOLIWERywOwojmVghpL6tAA2AAZ+EphuAEYADmGCyAoSdW4eABYE

kchJBEJlaSWh9YhrLvFUGcaIZihSNgBrBABhNnw2UnKAYnH+gGYVgDEeyCaXDYG7Ka5CDjER7PV4SN4AUXhvwSv2qAIgADNCPh8ABlWDdCSCDzoy7XO4AdQWkiWBzJtwQ+JghPQxLKB3BOw44RyaHGBzYcGBajGfMGZxKYOEcAAksReahcgBdA4Y8gZOXcDhCHEHQiQrDlXCDdHgyHc5gKornZonL6zAC+dIQCGIE3FEsgjBY7C4aD4swYTFYnAA

cpwxNxlgBOSbjBLRgDsX0TeuYABE0lBXdwMQQwgdNMJIfDghksgrlQchHBKtm3XzE5NljweF9BgkEinowciBwblqdfhe2wQTm0Hn8AXA5JQgAVLBQAAy+oHE/zCAKToK1sgZQkBDgs9q9QOttaCHaR26Bz6qC+0c9EFFqHGPCf82Ii39kzWM62OxQBMiZPteJxPvSdzQi87zph2ACCyyJuiQIglKEJQk8MFwr8uHjPC6boliOJMiyFxPOygaQQgV

JfjSaBfHSVwMqRJzkSSHLCFyPITAKQrYCK7pPuhsrynkKqBmquAag2qDarqgb6sQhoHuK4ymsWxAWlasxNPAdqOs646vmpBzeiGfqvmZwa+uGHCRmgIHjMsaaZsE9a5huhaaaW6SZNk4nVrWVTGeMTYtmFQwek+fZrnJQ4jmOsmTtO5yzswC6YMuq6eVOm7FNuxS7qUskQKQhDMDcJ7oueEhtB0xzoneD5Pi+XzLB+1JLL+BybNsux8gkoGdCyEH

MVBWGwugbyDL8kzVJMhGFsCoJmphMLvPC8Fzb8yGqtieIEmxbJukx5I0V1/pnSxR3lCdGl+JI2m8YGgrCrAQkHCJcqVhJ5xSTJg4KecSkqegxqDDwD3mjxaC7npLQMYZVEurJ4zigG5zmb6L1YzZYYRicIE8PyikZlmxkpQg3kYb55YBWgVaBjWdaheFJOJlFHq9jlaDycOgbPEluWpSU6WZdl/Yi/lYCFYUimlXolzVWe+kXleI1ILe3AtQcL4J

B1Byft+vA9f+/VAYNw2NWgY3ndBU0QG86YJPBkPqctaFrQ77y4XN8KTERB2sXdFGnVR40XXRtIR+dIdEmHD3cZauMlG9AkfWKwngqJv2quqCCanzCWKQad6HOKXzQ1psOoPD0Bq0jjRywIqNCYxgbY5wqdevjHB2Q5qDE5MrkU8lXmBkWtNlv5efM8FHmNs2HNc+KPNS8XwMlELdzj3lvXzouK4b6gVNbiMxX7ugcAvFAGIUWwKuBrV6D1WBTXcI

mj560sqxG5d9417m0AksL4f5zjv1ttdCaG04T9GqMsX48EPaTxWuhSEPscJfEREgoOJFboJw4rHBktETaYxKNReOrJE6cUes9PkfF3ovnRtnaUP1AqSQLkXeKW89xlyNOKZY1d6F110g3RG95kbnDCKFQR1kfTdwYZ3PuA8iaDDAaPdylMJ7nCniWGeFYOHnBZiFNG7NIrRRirzHhAtzg720fvGch8srHzimfAqF8FblCwJeIQUBfRPxtI3V+l4G

o3kDHecY4xv6BhfDwRM4CSjG3orwIavUAIDVfAkg4kDTjQIeJNd4nwfiIJQmg72hS4Q4JRGifa+DmTHRocQykADyECEjlQ9ilFzicierXUm5x06CSzl9HO7DGZ/RKADQusl+Z6n4apDswja71xfvaZuRk0bikSb3BRlkBklC7v3Qmn9Bgtk0QgRep8dElD0cQOms8jElBMVcsKy8LHRXXnFOZgtRy72lgfDKR9rHuNlp4kGpUMSo1QlVOoNVgllV

Ce/bWjkYnnBfEMHZEBkk6yAWlDJltXyJByZrbgdsGSYOmvAxM9xfxlK9ppSlTtfg8F+F8X4KD/rBwIdQoh0jI6kJSW0i4HSeUVFCJscOPSuJ9JTkowZ/FhkmVYTWcZipJmQGmdwn5IMFng3FP0ZZcrRGNARgZDZKNZGDH6PIiyPcgx7NUacmMFyrlUxpvovyhiJlBVZmY95nNLFfKBrY7efyHGiw2M4yWbiNznx3F4iQ2ARCBCyNUXAxBlBa2fgi

t+pKUWoGWO2H+/pYz/2jv6Ds6SLY6yxbk8lMDsLTTgq7JC9LVqMsqc2hCba6mHQaaHPlFCBUAIOe0uOYr7q0OTgqMdEAhmZ2VaMthYkfWcOkjMkN8zlLlwhntQMa0RGrOCesjxlqtmDFTMovZ9qjlOsch2G1ZM3KXIjdTSePkDEM3Vb60xwEA2rysSfHVYbhbrkcWlaNrjpbxqKom8G6Yqi4HgvgI8uB4USMOKQa4VAC3tixXEtJgYcVoDNvimta

A/6Bnrfkplzse37t0eUztsD0BXGsMwIUqagJ9s6VO5pUcyH5L4006VdD+mMIzswvFkoxmrp/euwGm9Q18J3QIwYgdaEw2NSB1uVqR7XrtfKw5KiTkPu+K6t9Hr7lfrnsYhebMANBsFtY3TEB7F70jRAG++pv35FNf501npiiDF0hqsAgXGhkdNc4VYYXZgavc6EKAjx9D6DUPWAACmwXzW6I64FIFAaoSl9TKDy+cTIxBiuQlK+V4dBWoDwWw2wC

gmwM11cgJVprOG2ulQKzh9EQQiwUCs04oFLiQVxrPXBiFRpENRAAEoVQHKrTD/WWsfwYlW2JOt0blpNtFjYBL3R1tJVAgTdHZrzUWu29B60m3QHIBwTjBX/J4P7WRfj/LzqCpjt9m6A7CHdJKL0kRc6F3SZVbnJ5mquGzJLrqtTizoxGoVG5mRF6eyGZxsZ3ZFl71D0vU+CqL63U3MBJ+r1fnEsvMcxFQNnyXPAYR6B/54GvM+e9YqXSkWQvrBC2

F/nvOwCHeKOjBI8XGiJfwMl1L6WZCumy7l5TTEGvVccJ0DrGBITq9qyr/LhXustd61rrrzXWshD6+bwb+Bhujcg+NmNMHpvy1mweebuBHiXAw2xdbuGIm4oIzrfYxGAGHakMd/0nYSU2zyRdrtTsXZuxJrdiprHHsca4293jk7RPDp+604Tueh2QFBxJ16irF0sOXaq+TTN/pw616DXdgxOUg80iI9HbcxR7ex4o3gtrbJmaJdGFsCRpiJjnSTse

AKP3Typ3Z55Dn/UtjbB2LsX9g367seGzz77zic78zz3SYBgun8F8f01IudnFESBL01SoEu9llwYeXWWctZC15cNXJXNfb5KJVrrn/jYqrobubibv/p1pCEbhbu1geNbgcENi1vbmLFBpNnlLBq7iUFfIcChmhtloVj7kaAgQHgMNkjtpWliiRq+PGNWiAmgNGKdrHg2gUunvRq2oxrcsxhhEyuxs9lnp/jnoDrysDuOiQoXgJiJiXmVDKmDpJkqt

XoGN9HXollqvDrwqUHqhXFEqjl/t3iZFEoPv3nOnesPiBM2JZnvtZg8lzvXkvn6v+vToBlviAb8mBtchBiUIfpWJfkFvzufg/kLiflEjfmAIwZLsUNLi/mlhlorh/oSujlEIVkAWVpAdrlVr/qka4fyg1jARAdkQAdAeAZbsQQNogbbsgVYWNhLNBuzjLHLJfH1ktpVAQTxjmmtiQecHeP0OQeilGLQaHhWjQVin1PQYWtbKNLRgnjNHNAtEtKgg

yjwQnnwS9txu9lIaISKgXkMcKpQsXpsWXsauDpXpDjXtDmug3hutqizqpmDNoVDFpjXDpjcRcPoejG+EYfsp8QTiBNGFPuTFolUbopTvTIvpALTivivM5nYq5i8R5rPgfvET4QFifmfqFoEb4Y0CEf4bFuiVLk/oLNEW/nEcrgUe0j/jVsAW5oAZkV/kkY1sUXAWSekXkSUfAWUYGEgSNkCagY7rUR4WEJgY0UaF7lAK0UQeyRtgWj0UHgMAMecN

QdicApklMEwZMfHmwUnu7Knixg9isQIW0VyvUp9nnmIS0jsUXsIV0lKu3uJkcfIVXjJpAMoWCZiI3mkc3gIuMFXI8Z3i8Rju6N6Z8beqZvZGomWs+jPnUdYbZjDhABCY4VCYzjCczhofCdGYGN4XkJiXzqiRfiiVfkqVfs4KqREWAFEZcHLrEcQErp/mkd/skbSWkTSZSVkYkbkYyeoSpiyZ2aUVKZyRUdyQibyTUegYKS7sKQeK0fCJgJeBKeDJ

0b0O6GiqMMHp1EMYoeRmMRMeBFMWwfAogsgjqUsenvqa9oIZJNylaV9vnuIUMR3P9ncBsTaaXrIeXgqkwp9EoXJq6WoU3locaM5LofWW8a3i5H3l8RBT8eLpYcORTvPqCXGQmUvE4dCdvLCWmbvnBd5kidmQWX4XmRifhViU6QLg/gSXYkSdWbWQkf6fSSkabjrk2cyQ2QyT1mycyWbuxUyVhhyRVoOSgVGnyWOfUeCtgX1jOZeL8G3Ktr7ouZAJ

El8PKauZRkRgqa0sSsqYSiTGqbuRqQ9m8NSrSgkMeRgssU9qsdnpecaY0tIdRL9ldJIWKpUMwJKknLKrOg6acd+Sur+e6cyZ6apPGMBSxaBcpXjjjlZFBWYVEsKtPoCdhXcjYdTr+q8uYgztzEzt8nCVhRmYib5siY0MLmifmUVSfs4L0aat6fflLkEcWa2KETpWWRWSlq/tRfEXSRSRrm2S8S2d1Z1WAdxV2YgUUUNX2f7vxXbjyUJaOSfKCg0f

BocNJa6DCh7vObxf2V0Z/CuZABiiBPtiklFHQZkj8LpWSnuQZUZXSp7B2ieXqRZQaesfsS+VsXeUJk5VaS5W5dOh5fahDl+ecC6XGX+R6QBa3oar6bXF3rIp8MGbjg6vjjFdErBXlbciCY8hcfYX+ihUmZlSmdlZhe4e6pmbhYzDmafv4XiZEXVTFkpaEeEURWVcWUdeVWvs1c/pWW1QrjWR1fWfRcxdSUxa2QNWxcbhxYLcQKyTxX7jblNdheLM

CnNVNmCgmm7mxmoOEkEphtmFlJtqgCHn0QxFQaOhpk+KMZkh1NoGdedo+awQZcUn8KZfdo7GeWsUISaXZSOhuZaR7QcW+faRXp+SMj5bXn5VccNaXEjvqmFCFdDVsmFHDVFXjI6jFTwE+iDACa+tNRAElbGZjeCcvomR8njehamd2emQKfviUDrYShAIAABygAKN6ADziYAGjKgAJXKADIZoAPRmgAcAaADsFoAPiugA5XKABfioAKbmgAIW5ESc

BQC4iEBGB2hPj3xZC/DSTYjMJniLgobbCWTuYIAYiGkmaFbmAEDb3KC71QCCjoh6BZC4D6hMDXEaEvDbD6gEASzlCN2t2d292D2j2T3oi4B+JsALbhDz0nDE0pkIAAASker42gGi1RitsanhpelA79EgEAQpi1zAWg/iUAmt1dCKNdetBtKlqA4eip7YZtsD/Qkw2gv4lijDc6NG+ljshl1QNK11Cxt1Zlp5D155h9mqV5vtL19lo6Pttlftdpnl

gdUmANsmvlwN/lbmgV0dmmB6HeUN/pYVBmydRmSdR9iNYZn83pD52BmdZOKDOd6NthNOhdONxdTp7mGF5duVldm9utEgn97d3d/dw949E9qAAAeng8EAADocAROAB7aoAEr6gAMSqAD6xoAIyagAv4qAAOpoAFfKgAsvKABxcqgHgbOKgIAIRWgA0eqAChioABcJgA2P+ABasYAI0pXdgA3z6AD7foAOV+gAy36AA55oADD/gAsHJd2ACqynXaU4

AJDGgAdSmABADNPVkHPQvcBKqDPavelvgBvc/FvUQOfeUMEAfeiN6P4u4GfRfVfQcDfVEPfaQI/d2c/f4G/YuB/c3T4z/f45PcE6EwgBE9E/E8k+k9k3kwU7gMU+U9U/U002010304M8MyU+M1MzkkAyA6wHM6jZAH2NA7A+MPA1iuQBQOg+gN49/X43/YEyExre85ExwLE4k6k5k7k/k6hoU6U5U7Uw0y0x0z0/00M6M5M+iArRNkrRgROYtX4g

AOJGDwQJJCKyXlDEMFqTC6NkPvjytzCjo7UR4Ub3gYvbYQJnZx6210YO2lI3V3a8H8Nu3WUfaSOiNe0mxzp7HXmmkyHSN/UnHyPOk/lKPh3/lR3aEo6Q3PEaEBk95Y56ORUmGhmDyT6GyRkJXIvWMIUY0Kb2YOEOMZVOOxRa4V0QPV13MSCAAGJKgGk4AHbGgAQjrTOz1gOgILMr1r0rP2o12HNbP72CMOr7On0bNHNwDX0z133cgXMR2DLlQ3P4

C4sQD5tFultwuX0IsVuxuoswPqsYsIMO6zXINebYsjuYOCtq3ipRDv1SsSAyukHkNKvPhLDR6DE2uqvm2EpvgYvW26u3mNpsPtTLDwQJDwhO0muZ4CNPX2ue3bE2sSODpSMzrOtB1Loh3nGJtTLKMvGqMVwPEaMYR+kBugXviJ3CqmHGNoAb4dicyaUZ2k6CVxueqIX53xn2NZLLxr6djdguFuaZvk7QA5voBjslvBN1i4BziksfMcA0XOD6jOD+

IZCoCAAyEYAKMGETgARLGACFSqgAoKgIANFygAb3LyeoCABnyoAN4+RbETgAsJqADC5oAIAMgAn9qAAK+YZ4AN3KgAVOZj3pMROADgFoAH7egAcHKAC10YAMN+CggAMVmACD0QoIANK2gAq9GADgSoAGBKZbszJw4VmIizNbqzNo6zO9jbOz8irb+ADbB7xzgYpzPbD9/bacg7r9w7LHo7Bb7HQTnH3H+DZLET/HgnwnCAYnknHAsnanKnanWnOn

HABnJn5n1ntnaTDnLnHn3nfnQXYXAD8LoDSL7jTOaLC7mLHIaDJXbHxbHHSGVX4T5LdXLgDXTX0ncnCn7XCnnXhbenRnZnlnNndnHATnbnnnvnAXIX4XgKK7zuKtM24l5QmAAAGgAKrMjVCZZCDrWHtbWUbocUGvhUbqU7HFpaW7baBaslAsN6vTEtqIScGAjcG8P3XftmtGkWvAdWsAcpK2uip/sge/Xw3/XB2A3uvkcg0BVg0p5+to7aOyIkyJ

1hsp3YdDzxjgWEdRmzfAnxu2OpV0641psuOJRs4i/ZuePoCACGJHS2hoCxFzO7wGY5qrF8s/Fwr41u28l823syfel0b5l52yc92+c5cwKIVxwLc4rxACr/8+r5O8A9N+A0x3O+i4t4g3y6u1XagziyV5ux91gXuKVPBHONgAALL9AADSAAmsK6D3cwWi2K1BMKq9QWvsdYSkMNoJVZAKj4+3bWwwkNUF8D8LUtw8a+Zfj1ZYT8+aSNa2T0B0Di9Y

cTIx+XI3Two6HR60psz968aK2LHRz1sq2Nz98TFf0P8UR9nbnQvkhZR28qhcmaXQTa40TUx8QxIIAJ5OgAgfqoCABJhKgIACEZgA+OaABspoAFNuangAScaACwKhEwAPJLi4iGeACIFqU4Z4AEfbQAGV6gAM21AABnIRNvmaTQABTqgATfjAAM4kedSmGvGbtF2XpQAlm69Otol02YSBtmJvJgGlwy5sYsu5wHLrb3y6QBrmRXEdif3P5X87+j/B

Tq/w/5f9f+//YAeAMgHUs4BiA9zsgI97TsZuWbFFvfXnZjFF2IxNAvyzXbLdne4fBatuwWzKBqgcffQPCGUBlZ92bGDPke2WAw8yG+GA6sBBGKwMUw9DRhpYMF4o8dWLBOjH7D9iftG+/BH9u7UtZt9Se9qO1iI3cpyFZGChJxkDUZ6wcNC8Hcfj6SQ7aZ2eqHTntrwRqhs5+fPSfLQxRry94KpHBNnYQLrJsqOm/EuqILLqy9iOB/dAIAAN5QAM

r6gAWSVeBgALH/P+uIegQkwiZJMUBUXawTr2rZ69sBWUYgRADEBZAmAuzQgWb16GX0re2XG3r2zt6vQHeTvWuhUOqEIC6hX/RoS0MEFe9sKvvBbkuxHJINsK67MPlgyUHJ9sAMAAAIrMBBgp4domxDB5LkIe2fBiP0GNpw9qG6rJSkj3vZ2DpiV2OYk4L4ZN8LyLfZ6h4Leod8PqPgn6n4L74BCocaqLIW6U9ag0x+kMSVpEKeLRDuygbEyCsFn7

RUkhYULFPFSzqJUbGKVeeDkI35S8gMO/IodnRKEQBAARiSoBAAcxmABLNIEGSQZ6kXStlyM6FYD4a9bC3ugHwFDDj6BzYUdAFIElByBUwygfOlmHFdnezI9kZyIgRTdEW3vKxlsIkH+9l2ew2NgcPkFHCvudUZbHuxuHStdB4PQtJD0NrjET2efeHluUyQ8BJgXwK2jHnVJo82CKwKvgkHuD/C8eLggnlMmEbuD8kDlAxmaUZAgioR75NOC6wH5u

tFGwQpEaPzuLj8TKbPPQpzyxRHIQyvPCNlMDioWNiOK/MjtB2yHY1ch1I+jjlT35WMGRrvelgC1KbBNDws4TbjVw4B0CL+N/B/s/zf4cB6h7AkpoANAEQCOAUA3gUgJKYRMgmnYrjouF44LDah9Q1YQuOTTYZ/IEsXjiqI5ElNgm5USqN2N46AAWs0ABM6Wp0ADyyoAE7TQAHe6anQ/oACB9dkRE1E6AAsOUM6AAqOUACFAYZ0AAG5oADFzZzlE0

AAlJoADK/QAPrmrQiYFinQGYDa2gonAbvVFGpcRhkosYV21voUCtc1Ax3kqNroti1e7Yxca2LPHks+xDAwccwOHGji/+44zgVOJnEIC5xHYiiSuPJZrilhG4i/sk2CbbjuMe48lgeLIknibglEiJleNvGPjnxb4tkR+O/H/igJoEiCTBMm5TsNhs7MQX7x2Eh8R2JEhlkePIloYpJvY0/v2MYFDjWB3/RiROK4HTieBbE/gSZKXEiSImPE+AcsIa

H8SkmgklNLuK4kRMxJJkiSeZJkkKd7xT4hTq+PfEcAvxv4gCSBLAlQTYJr3A0ZXRNFR9ygC4AAGr4BMs1wMYNoOY661M+doshgkGeHGD/QzosWGYM1ZfCLqbDaoJemQSIcmMixXHi7VNbN8wxNlYnqCPNKAcIREYpDqBxp5JiIO9PVMVWMREj8VGLPCGuiJQ5Yi0O6dQxgkPxHFjJgwbcxkv1JFi9yRSbGsVSMcY0iM2bjEQWVNrqABjElQCAAEI

0AA05mqKmTcjNebYKthgLi7dDDeSXPAU2zFFECsJ0oyALKLy74TFRI7B6S9Leml8NRmvG6TqJVJ6jdhgffYXINroKCxKOUpNGwEuAWitatw60fcNtGPD9aGmWqVrzeFjE4wXovSj6MuouwlKgYo1mnmDGWUgRA0onl32GmCZwRttVvvGIDowjHScIlQvnHTFLSUR8SSfjEOn5XoQ2xhRIcWOjCbS9wZY5fmSNdLIVaxF0+sYTTl43SGRsM16SZKV

hQBzJgAY3lAAi4lqdAADsqABGoMABwXmp0ADvcg+IiaAAmxUADfioZ0ACLdoAE6HQzoAE7tJus50ABGxoAGkjOCYNG+lIT9eTQVCcb2BmYSAZJA8YWQMmGQy0iBEuYeUDNlkTLZNs+2Qp2dluyFOnsn2f7ODlhyI5MczSZ701GbDdJ2wrFljMLlPTzZgkgmVbJ47ks7Zjs12R7K9kcA/ZgckOeHKjmxyMpGMuotlJKjlAEgPALQJeHoBx90+5Uo9

pMEqm7UlgJfbFK0ijYujCUHw5HqX1sEtT3gB5JBG3mx7dTna0rPqdzKEaDS+ZkY8RmNKGkize+iY8DpuUH5QcERTPGWZmMhjqMxMUQ3MdPxPYFj4aWHCNu+DiHEjLGXmCsZkLsaUj0qzhLKldMbFeYGRgAExJUAgAEb9AAECqAAQ80ABfenHK14JzfpKEnoZKPQnXoQZGcqUVnJlE5y+2UMl+oRJHYkKKFNCpuUIK1FeYUZN7NGTNUyk3SjR2Mxe

TgQABSWUXAHAEvoftSpdwhStwF3kUyfgwqPPifIakLsqZ1GK+aw3eAxgvgkwfoPMS6k8Mn5B7F+c22Ii8yRCJPMEV4Ip6QiJp1PaMfOmmmAKUxQ/NMYtLg4s9fWq0rRgrPdA8B9pEVFWTtKJgkxicWso6RkPF4UizpOCtCgUNpFuFjZ+/EroAABzQAOSaLdAcffwiaAAkIlZFsjAAlwaAAtf1qWoBO6L0iJoADc9QAId2gAZtjAAy3JZNDOiUwAI

K2gACqVAA5AYwSImgAZPiR6NQwAHdu9nfToAFnEuhbDT5E/SuhTC/6bgJFFAyMJEojhdhOt64S5RfCodiO3KWVLGBrS9kc0taXtLnpXSvpYMuGXfjxlUy6CbMvmVLLVloi7SWkOcbchxBqM/SVIGkFB8luofY0Vu1NHoAzhRgdQZgGWBKKfuW85tpElbD6KL5R8jcle0amfCGZ51SxXCANb/B2ZupXqYCNcXhif5AmKMeTwnSU9u+/tP+VQKCWBC

Ge800BREpRERCoFGImBe6HqmJLLImHcNicGQVgJW8XwHoqkJukYKslp0tKtR3wwb4ElwKgpTvgIXB9bp5QQAKYkqAQAOZGoneGTFxmafS4hiExhQEqFEcLWFyddhXss4U4Szm5yvOdDJK5GrTV5qwBlpJbk6SQVekqQcJRkF6r5F5QHGarXhUQB7gAALRr4IAvgNwLgFotJk6K+Q2KktPeBXnUzoktMzJEX0Pll8YxdGF9m+00X18OZ1KkMf1Lfn

uLrS/MxlZ3w8W+CExHKgBVyrmkgKQh3ZMIeonvmOtoFIFWREpTxHKzjkSQ1sO0NKDpLY2Sqk6VjVVV5DpehQwpcUJK5/KVl9AwAEXRAzCJoADGjAZoZ0AAKae7MM6AA300ABc5oZ0elt0flNQwAPjugAPO1AAgMb9LnOgABudbZ6yhCbrwFF2qU5gMlLmwvTkuqTlEws5bnOZL5yiJ5QbdXuoPUcBj1Z6i9TervUPqOAcy59e+s/U/qAVgaoFZIo

mDSKIVYaqFQek7kYNFFpUUcPGswCDAMQC2AAFYYq9aeinNf0DdHUyVgha7SkrJsHMFr5cIV2N8FoZBja1XM2le/LbWfzva38j+X4uhH/z++M0oBfCNUJ9rt04Cn4PLPWljqJVN6BBZKs/hthF+wvRVTrLX7YKnMW/fJfgqKVNiSugAMxJUAMTQAAdqhnQAItugAcuNnOgAFA9AAbl6ABZ5UAARmXQr3kWqtlgGudPapdWOqj6zqjtm6ty68LPV/C

guRIDc2eafN/m4LeFsI1Iyfebc3UeCt5ZO5DR1G9ANGs+54z0AC2BbBiC0hQAKAWgy0QewzUQA7wnGqHj0XXJkJjFR2d4YmAsGWDLEs60ta9SfbvAxNcqyBVwUflfs61r8zEHSqU220W1imuTcpo7WBKu1EssOuEtCFg0a++mzZCKriHwKAliClJTYoVVMdF1us9frkvs1arHNm653i9MqGAA4FVon1LIpqAJJoAAlTQAFPRgAWtMIm4WwAAxKRb

Qzv80M7sjnOL09ZbOptXbKgNzCh1QcrA1HKINYMvoTwumEDtMt8GiQN9r+1MCAd14hTiDoh1Q6wtsOwtvDtbGI62RyO56UVuEElbg17cueZVqBWRqaNcK+rXvSzSQh2N0pCmU2BeE2sa+/GpYHEEYLjboocQqbdRH1bfBHalKu6lJsepuD6Vm2r+ULLjG7bRZqm2EWcU01Szjt/a07WiMFVrSLtYodqBOq2kExp1BgzWYdIXXWbyOes86am0ulpF

GOzm53qU2SaABmxUABZ2oABwCfuoAH2jchU50AC4BFUwoVLKY98evuoADXlQALBBD/NnXExelp7TVmeuPf3UAAvZoAH75QADnagABnV71KeiZoAEg5QAKAB/TQADAqddH8UsthabKeRjYBhRjri3Ab9loGp1eBpS2nL3VMGtzHBpHYR6kmWexPcnsc5p6M99nFfbnoL338i9JeqpmXq30V6+6NehvU3tb0d6u63e3vfZ373qiA1xW7UaVrBWhq3u

VWmFQooKjgA/ohwOAHAHxAhRuA1oaAJsAyBbMa0IwBgIQAQAUB00S26YqiCQS/AVpMowKVkBlDZh9A+Ie2NMXGAIB8D+BnoH0PQNQBMD6QeA44uW3SbiDQk/yOQf0C/B1tO2tAzuIwNYGcDYIxPggBfAQAMgjgIQPoFoOkGGDnBu4HlIKxWBNA2IJVITuewdBD6JBtg2QY4M+LxprB7jAwYWxsrU4ShzQ1gff6cqJQeh+g1gd+AAbkJqARJCYfYP

pBzDlqmbpVJsMqH0gEsXoYlucOiH6SUtfLp4awPwhRqYtaWouT8PpAYCc4XNJpGEPKGGDvwAuNoZZAgYLg2Aa4DiHRVoBnAbYbQLQ27CJhEg6st0QmCgPMAUjTwfAMnx1iJhRt+g54T8AX7ekE6BQCAEYDYAGBgDncAgEIHAhW12oe08YJgVCP6BtDyHWuMOptIQAwQJAQfbwGMOTHiA+IBAJ2xw5QG5jcfNgMpACO4BNAwQHkhMfKhOKTUJQWoP

gFKikBlAQIAABSthUwvALnq+GjCMRTgVtAAJTogQGygHUAVlaDnHcAVxjRDTOoC8BDCTxhIK8YGN0GZmkcQw/4k4CYioD0yEBgaHKjAFiomQLYzse4BXAujJzIgEsdQBYm9VjvcA2gAJMCg/EqLTE6QGxOBh9ABWO4KQFDAbpKT1J84LSdID0nNj2x4yASYGN2AWNoSZgLiEd5wA1jGxx3lyd2PAh/EjAOcK0fwDtHiZd0NIO0BxgnMhAlwAwBEY

kQMdrpTHNUAYFxDKmYTlkZGclnggqmZTcpkNAMccDMB0TBSAYVlDj6ZAQeqNHOpkSKn70mAmQQeNrntPjGKoKRTkxiZJNUmq6c6uPiQEPxCm1FcAUrMGe5NhmoDRYTAIaeCAqnOAoptiLKM3aFQ1twQK0A6BAAOggAA=
```
%%