---
excalidraw-plugin: parsed
tags: [excalidraw, quants, learning-path, regression, lightgbm, xgboost]
---
==⚠  Switch to EXCALIDRAW VIEW in the MORE OPTIONS menu of this document. ⚠==

# Excalidraw Data

## Text Elements
后续学习导航：从统计直觉走到可运行策略 ^title

主线先建立可靠基线，再升级模型；不要一开始就跳到 Transformer，也不要把“预测准确”误当成“可赚钱”。 ^subtitle

后续学习导航：从统计直觉走到可运行策略 ^title

你现在在这里 ^wquZIli3

可选支线：深度学习 → Transformer → Foundation Model ^GeL6zYIg

进入条件：树模型与成本后基线已经稳定
用途：序列、多模态、文本与迁移学习；不是自动更优 ^42pssfzJ

每完成一步都应留下：可复现数据快照 + 代码 + 图 + 样本外结论 + 下一步问题。路线的单位不是“读完一章”，而是“获得一种可验证能力”。 ^foot

0 数据与收益率 ^s0Title

为什么：先保证口径正确 ^s0Why

能力：复权、对齐、收益率、标签 ^s0Gain

1 OLS 与损失函数 ^s1Title

为什么：建立统计直觉 ^s1Why

能力：散点、Beta、相关、R²、残差 ^s1Gain

2 因子检验 ^s2Title

为什么：防止样本内自嗨 ^s2Why

能力：IC、分层、滚动、泄漏检查 ^s2Gain

3 树模型基线 ^s3Title

为什么：学习非线性与交互 ^s3Why

能力：LightGBM / XGBoost、CV ^s3Gain

4 风险·成本·组合 ^s4Title

为什么：预测不等于持仓 ^s4Why

能力：约束优化、换手、容量 ^s4Gain

5 回测与执行 ^s5Title

为什么：目标权重要能成交 ^s5Why

能力：订单、滑点、成交、对账 ^s5Gain

6 归因与迭代 ^s6Title

为什么：知道收益从哪来 ^s6Why

能力：监控、失效诊断、反馈闭环 ^s6Gain

%%
## Drawing
```compressed-json
N4KAkARALgngDgUwgLgAQQQDwMYEMA2AlgCYBOuA7hADTgQBuCpAzoQPYB2KqATLZMzYBXUtiRoIACyhQ4zZAHoFAc0JRJQgEYA6bGwC2CgF7N6hbEcK4OCtptbErHALRY8RMpWdx8Q1TdIEfARcZgRmBShcZQUebR4AVniaOiCEfQQOKGZuAG1wMFAwYogSbghmAAYAITZMFOLIWERywOwojmVghpL6tAAWBP4SmG4eSqGCyAoSdW4ARgBOScap

BEJlaTH5yuHIay7xVF2piqhSNgBrBABhNnw2UnKAYnmANgBmHn6AMR7ITS4bCXZQXIQcYh3B5PCTPACicJ+CR+1X+EAAZoR8PgAMqwboSQQeNHMc5XBAAdVmkjGezOF2ueJgBPQRLKdLBmw44RyaHmdLYcCBalGfMqJ1WoOEcAAksReahcgBdOno8gZOXcDhCbF0wgQrDlTTitFgiHc5gKoqrZpHD5TAC+dLCCGI3A+4olJUYLHYXDQfFOPtYnAA

cpwxNx+h95gAOSofD6xlYlQjMAAiaSgru46IIYTpmmEELhwQyWQVyrpQjgxFw2bdfIA7LH+jweB7PrH5vzTkQOJctTr8HSHsCc2g8/gC6dJKEACpYKAAGX1g8n+YQBSdBWtkDKhPmtXqdNtrQQ7QO3TpfV4bxTkFFvAmdJmxDmfOWdMk602UG2Xr7J0LKAfS5JQo8LzphMACC/RNmigLAlK4KQvckGwj8WHzHC6Zopi2JMiyFT3OypykgylLUu6z

pkoy+JHCRxIcsIXI8gsApCtgIoLOKdIobK8p5CqpxqrgGqNqg2q6qc+rEIaEjGpU8ymsWxAWlaUxNPAdqOs6CATqgHrKXSwZ+hxQZMCGHDhhwkZoImlTtnqGZZoZU4zqsRaoaW6SZNkwnVrW9aGfMLZtmFbyenxfZrkOMmrGO1ySR5CDfguS6rgOuabtuwx7qUkkVDwx5omeEhtB0hxore/TJnST7jA+EBvh+qBLM1P4bFsAY7HSV5HKBFHgehML

oM8lQ/LG1SxnhhZAiCZpodCLxwjBU0/AhqpYriDHlGybq0ZRVLvjSAZHeSRGMQdql+JIGkWasgrCrAvGgQJcqViJqxiRJ8UjrJBq3hASk8Ld5rsWge7aS0Dl6eRBmScZgarGZnCPd6Vl+rZ9lGR6iYuZmwQNjl05pac3klmW/lfUFdYk82rY8JF0Wgf265ScOo5sOOKWbulzCLpgK5xRuZN5buslFcwHylaeOnnpewFIDe3BNi+pyNRrqytWd7Vf

rOv49bwfWnAN3BDXRtyjVBCQwU5KnzchS0QWNEDPFhU1wrG+E7Vd+2kYd5FWydbUoyUw30cy12B7dbGWhjkDPdxr1iu9YKCbTonqggmpoNJAOrHJCnoEpHzg+pkOoND0AK3DjQ7qsLpI+KHymVj6N8u3vphhGdoev0/SE25fNk4Wam+eWAVoFWpw1vToXhczTZRdFo6i5zCUlEl7n87OGXC1lHOpRLxQFQerL9HLpzleglXm6raCLIkDVjNrJS6w

sBurF1f4Af1ysLYXWuK7F4bxqi/Bgo7CmC0UIQlAZhD4CIfgwV9oRPahJY7AKoqdWkwdKL+0wcxU4nJ7pV17E9LiPE078Qzp9QK2dxK50kgXPUQMjTin6BXB6UMtK11hkZeGTdEbuk4d3ayicGAdxsn3d0HoEjNTTETBADNUCpXHj5amFYGGrHniFSSYUmYszXrFbK+cuZ9h5slUmnkShzkFplDeJ9iiN0KFLfaCRr42jrnfC8VVrynFvD2QYr8A

zv2mNRT8nUjb/l6qBc2aBLaUQQeNZeMFwGIVgS7G2sI4Q3E2j8cu210HRwDsQpuIdIm8GwYQ1kWCSGsTIQnLupxk7UPajFSUdChIz2+iUX6zD/psPksDJSCRuFVxrrfe0Dd9KGWMs1NG/p2riOxrIhynY3jD2JrvMeFMJ5aOnoqPpkA9GqMMRFFerN15mM3oXbeVjdm2MgPYoWItbnOLAK48+0twF1DKj4iA99AGP3ah8e8oTnzNU/lE78MT/5m0

AYk7BKT3bvC+L8TJzs1KovhIiZEqJim7VKUQsiFTjpVPDgIK2tSmJkpKKQnhKzWlUNTh09O0p6G9NVDnPOdzhklxBuKN4EzmnVz4dMoREcREOWFaszuzLUbSJxv3SoTYmzbJUU88mXkDl+W0dyuewVzlL2MZ6G5HNWGWN5jYnVdiD7vOPrlFx+V3ESFwJUAFAiIC4FIBcKgoLvigSfDGUCMLeDRO6rE44ADqrIvwSNFasJoJ23glixaOKcnoHONY

ZgQpAhZDQcS4iN1sGhz1nEZqkcEC0tLQ0u6TKKElDaWynYHKaxcuOTyphfKrVF3YYpcUW160QzFX26V8yh3yuWU2yASyVVyMqLGWMmrVHqP2Zo/VRzZ66ONYvIxVyTGJQ3uOyAO9R7PIgHANg+pt1aXyI0YogEn1aROWAB9j6X5aTAJUV9UwTkQHwKEKAdx9D6DUA2AACjewt5it7Ut9VAaocl9TKCGacTIxBkMQlQ+hipiGYJ+rYBQH8uAWEWNW

Jhwj/rSNFV9f6tEQQiwUG1QLN5R9bWnzcUXOj0DvHevo8RmqUYoqQtDa+Slka/7xtWAkmNCaQFZvdpNaas101wOWhhbN5AOB5t9f5Ittb6nkvJOWsY2gq00owXU8pDLGmNs4i9J8bbaGcp6V2xhf04P3P3AO0u4ofYsVQky09FQZVGQC9OyR871l40qIsRYq7WMbqplurOu6F4GNNYe81pjLUUYeTasWl7r23srPe79z6f2vr2O+79X7H3VcfUqf

9o5gOgfAzIV00Hb14YjlEUgSGUOdD65ATD2HHAje87RAjRGSMhHI/BjAEJqPEdo+UQTAaMP4GY8ln+DqOPFa3C6yWPGNtg3lgJubwm0BvHjGJ02OtJNwqjUAxFcb5MmcU0m8a6Y7YO3U9kn70AdN6YLf+IlRnbPUopbggMFmanWbpUHVYjLyGOZTs5zpJQPruZ3f03li2fOlD80K+LoqFShebqI+LUWWlKp7jIuy/ceygSUSPW1GjUtT3SyUM5+7

Lmr1y8e25oXz2c9OKVg1ioKuNaq7+5rtWP2NAa4+hXjQWsa7a6SDrEHuswejVTgbQ2cNTf5RhiEE3cPTeDrNmjC3RvLeIKt+bZGNvXbpEx4je37UOMPk451XzXVnfdUUm+gLNs3dQIsbHj53SPY/s9w2r2ZMlDk0kxNWn3bgMgXxkoSEM2oVRTm3T+aDOQ6R3Wr7OCw4I4UzWyvxm7MNvRyypzb1XMdrxwBgZvaCu+ZGRwvPkAlohf72FydCe53S

LwQz6yC6xT9Bj4sVsGrZKuR2Reu1AI9U850XzvdWWmYdkqF2HsFrHfi6O3SKXd7H3K6fbVprWvZefofC+5rrW+ztYMJ1qDBvHdSREMrczdQtxths0Mbd8NBsXd1soCSgqM5s4D0BI9Pcdtvct82NHEPlA9vk3UUCuFLtGJUDAkFh5hxgHsw0k8f54VU8gIPsM9vss9ng0kMknZC94ElMS8wdy9RI/ZG9ocwJrgzN4dLMCEBD6UR97NW9KF28aFTh

cdedIBe8icBVRllJh8gU1Ix8ltqdZVz9LJGdotlVYtEx5gvgktMCUtiBJ4aZ99TlD8FhsshdY9AMT1x8r81E95Vhb9yt79Ksn91dig30H8wBVdGggiwBNdgjtcQNf89diAetYNzdoCTdJtICUiEDLcIDADjdYCHd4CxsVskCCiUCPdttdsrD9s/dHVOMTsz58CfVxkiD3cGNQVzD+hg148qC4cI0XtpNPs08kVBiYdM83ZXhPhvg/h2CNNi9Qcy9

C0K8SUbNJChCa8K069q8odVi0cxVZ0IAW0sd21M57CMRCdHdi51DyCKdADwsPRyC6dFVMZGcF88Z5hoxLCJddVN099DUMt9EnCD0XC2Z3CltPD10fCDc/DGhQj5catX8Vd39n9giv9Eof8wMEikjDdx8gDBsQCMiwDsjTcCScS8iSi3dCind8iKSyi2iKiMCvjfd2MA9xZ6juNUwipBQoA/QCAvEShb4gU/EH5SDbsEgui0AEgPhoUqloxQJf5jZ

zDmp08UUlNnhwEmwbhkxAdM1gcPYeBCkfhNCCJi0Y5BDq0RDql69aU6xmAfwUdm944FR9jDiO8FDuklCzie1VDAZB9B1zCbjeFH1+FdJZkEZJ825DCJF6dnj59TCYwEgh519lE11vD89d87C/iD9MtATBdrk8tL9HkqimTsCnVWSg9TsOSNsAB5OAbkzgAgGULgFo91co1YW8eMKlCAJ8foeYaU3o0TZPAY8g+JYYxg62XU9UzU5omBbFIvLg+Y/

TRYvgkpEtJvUY4QylRHZYiAG0u0uOJpJ0jHdpFzN0tzD0lQi40nY0d4gMzIgQO45SRMufcyaM6fF4uM94RRDfLVIsnfH4jMjzf4k1IEvMkXfLMEwsxkyAXwvIBEx/SreE/wxrAcxrYcv9F/NEnXeIrrRIgAyk3EtI63O8p3fE3Iu3NbUowk53ckujVshA9AljX8qQA7FksILjH5as2snk/AKsoQCHcPK7Okts3ie7TWbgKKTs8NT8/ohU2MNfWTU

clUic6oDUrUmYoHLPbghY/in6fg7cqvfrWHMOLc4iXcnMILR0yRF0+QrpM804i8yky4ofac1HbQquKnB83sx4/YmLZnORSKT46/aw2w6XfHBw7MxmXMo9beUE4ncE1M6CqE2CpCtXQIxCmE79aS79ZwOMJsdCmI7/LCjEnCrEsivEnIyk8A4ksqqAak70yjYo+3Gkn1OisbBin3F5FinAssvAkPdAbAEQcHaoMjZQFWASxiYFaqQNeqMShyGap7X

o3sqTY2D0WNECJS5glNOCYdLyLJHUza2CNNJY1cs0ypRaky00nY6QvYo81tVwxQ+y84xyq85SEVILUdTSIMyVUM4RSfLZSMl8p4t82MvyjZXKwKrwvZb47nACsKiAfnI/KK4XGK0XDwyCoK6o5k7qtitkjiiQbkqAAJfjRibMYWKPEJWaoyFdU4KS8YOU2g1ATozYoYhgja8Y9FKY7Uuc4HLSxcnS/pPSk61Y80qpfY6tbY+0qQlvG6tvTHV02yr

vc8p64ipyv0na5vD624yfeSmMwGnykw0Gt4lfCGiEtM/80KgDBGnM5eYEi/Sk+KqG/kpccoQAOBVABbv0ADNtQAAzlAAf7UACogwALH/AA5uUAH2/QAQujAAXt0AEnIwAB1jAAGJUAHvlQABfjAAZCMADVvQAU1d8JOAoAcRCAjA7RQJ0Qc6fhxIsRnNTwlwYIiBlBllAMEB0R+a3zuT3Bq6Ng66oBBQ0Q9AshcB9QmA+8ltHgNh9QCA3lXbPbfb

A7Q7I7Y7E7U7M60RcA+K2AAAlcIfOo4U2s9fugACQZvmG0ClKwP92xu3yBUoHHokAgHYsaOYC0AJqJv5MBVJqbogFqmaifBjwk0WsGHppT1QDu20GTFZlAf2OVPr1xUnLUpnI4M0zdl5vB0MwkMlrWItLFqs30rXK0OlsPNluPPuvdMeq9MvN9P8zjFvI8sn2pufIVX1vfMNsTHGESyTI5wxrNphotrpgBMipttApRvArivRshsvVfonu9v9uDvD

ujvjuTvToztQAAD1H6EAAAdDgdRwAbjlAB+v0AAlFQAL71ABprwTsAAN0wAL/VtHAAYf8AFjFQAcWVAByv0AEIrQAaPVABsf8AFg5QAQMjAAAOUAAB9QAac1ABGHUAGfYuO1AecUHYu0gDIUgSxwAfTkvHAAoo0ABwCQAEIzABpW0ADHFQAOw9ABcAkAHvowAZX1AAEI2SYTsAC1YwARpTcnAAgBmzqyDzoLqcNVBLrLvwArpvirprrruCEbrRB9

BboIDbtrvKE7rgG7pzr7u5FIEHuJ2Hv8DHudokHdokenukbnrkczqUZUfUa0b0aMdMYsZsYcZcY8Z8YCZCbCYidzSiZifiaSbSaybyaKdKYqeqbqf6hXvXtYCafYZ3u5H3oAcPuPpIUvqWfQBWanqkdntkYXoUeUbUGCF2Y4B0YMeMfMasbsacbca8b8aCdCfCciceDuYSc8ZSYyZyYKZKbKaqdqbRFeRLLqPLIaL6pagAEchAAAtGUIgMPYm0Z5

20FYyMTb+mm0W8ItYIF2MbQGPUB0BzsiB6vXFT2T2LmzgnmhcpB46y61BkW86q0lB/chzfBu644ztOGhylWl6pYShnEzylh2hmdadV4phig1hzfKCkGdMrho1CK9qZw/h/5wR7mIrER8+sRiQQAA3lAAG50AAp1WNwATfjAAZxPqdzs3oWCfP6VafA3ackVfuGbrrECyCYH6aYEGfwALdGa7rpB7qiH7pmfqubVIBHo4EWbJqjbjcTZTc+c7u+fT

b+bcIBYPqPs6i6tLMvXIAoCvvQBvtxsaIAHEEBlw3gjAABNGUNDZs7NQVkUwB0S1YJ8BRGhxPXoj4dWGS6NBIYFuV0BzN+g9ayB1Uu2T4O7NV+B0ZzV3g3SlcnVkkM6tqDB8QrBwQ3YvB2QuWmynHIhzM5Q5W0LVW/zC7EdSuMdO1ydN1x12fXW3uRh8FFsRISUk2hKr1827dS2xw3hk/M/fY9mAs0N7e6AcFiABOwASATAB6U20YDsAEfbQAMr1

vbUBAAkwkJZueJaYEE/UZ+DUnrD9FQAAFk2B5J8BU3GmjglgWmshS6c2OmbQun27yhem36BnzAhnumq3xma3Jn63ZmBRm2Fn8Bp3mP2POPeP+OhPrndNbmxOhPJPUJpPOA5OFOggl6vmN7fmw2bkEBAWhyR2T7aiB3J2HPZ3mX2T9wio2w5BmB0QjAAApL1Emnd4S27fdkYcS1w8NM9/+gYs94Bm91mO9n1RSx93U598FQLWB2Y+c3NbS5B4D4W/

9vWQDy6Q1iyg8qy1lI4zvE4mDz0rzK1shoVZmW13Qh85mby51uM74Ts9nD1gdymGww5D0q23hs1Vw2j+24RhjiN9AQAbfjABTRUAENzQAN7kA7ABECxccADg5YpwAGnMXaLHAAn3UAHm/QAZi9AAs7XUcAApXQAFASA7AB4vUAHQlQAQAZAAsTUccAEADBHwAcNMvv3vABB+MAG/Pb2jxwAejNAArwMAAqlQAF7NAAMeWU/7fanq+Lo07ae06duF

krYkAM9LcG2M4rdM/xurdOFramYHsbaTls9Hvs6Y7u6e9e4+++9++0cB9B4h+h/h+R7R8x+x/x8J/cdJ8p5p57bXtC63uI/7Ei+HZBcxsZfi7BfbZndvtZeLrYDfoFLEdBQpoPbVhQpPYA7qkq+NnbBlbFNq9bjWsGjZpeA5sxXUv2oQc/aXO/ZNLKT66MoG4uuT9QdA7G7kPZUm/NZ7zg/HwQ4W6Q9cuC3crQ5bmcgBrofW9w8SAjKLm/JTMdr/

M4bI+4eAqRtO9ipDesQHau4gEAHnrQAGO1invHABTa0AF+EwAFL1ABNV0AGg5AOhOwAWE1o3AAHU0ADtjQAa/1ABzR1QAAGpUBABjuUAEAPQ/1AQAP7UL/AB2Cy+8ADRNQAZb9AAu6Iv4X8n8ADvUwADIyanAB72O0cAAhboAFVlQALLyuvZJoAG/okft40AAHXrk0saAAYAKJ7JNAA7cGAB0/W8aABZzwTqAAYrMACD0YAF+AwANlKdLdTmmzC5

hQyBmncunm104jMOeDdQzmWx57s9s0AvVYELys6i8Di4vVtpLzt7D8x+k/Wfov2X5r8t+e/C/mfwv7X8j+d/J/q/yP7v8J+3/P/oANAHgCoBsA+AUgNQEYDsB+A4gaQLNghcfmJvVvoO3N5AsYu+8GoodnC6gsp2THJLr1UrKEhKg84RFmNX5b40CuvQcSvsUairUxWi1T4BezGBJB7wIfcUAq0a5KtVSUfaYu1w0px8uufNHrkLV1b9dJE4tYbv

WksqvkDi43eWlBzsrTdLW8HF6h2CW7E49CEWDsGtxr5M5cYTDN4PV224/lPWe3EKh319Y8N/WIFaKkGzo798HBOnAQZUFQBb93ugANlNAAU26ABw51p4UCHWWbJnlp1oFs8+e/VfyCW27jltWB0AdgSUE4HTNrOrSXgW22jQQBJh0w+YUsMN59swuDHM3lFwVI2Crep9cdufQS7OCHebg1kJUApCSBRgW7RjmTVBRvBAhNOH+gBzCGDlZKMraITE

LD5vZ4hupD4IPBggJA4Qb7OYmkK1bLkk+pKLIanxyGYNMhRrGQs2mKGQdIAD1coYXyWzF9jQ3wGoXMir71clkxhBhi0K+BNgvyyZDqiR3b6HcKOAw7viCVRoQV6OxHQfoAC45QAANygACTkA6ujQALvyeAwAMfKgAEP1AAxtbZNlhqnBntmxoGFD822wiAEW2zBPB9hLAy0WMwma90uBjueZhLwc5KjVRGo7UfqMNGPDjenrV4Rb1HZ2DWKPw23t

cJcHB4ARFQSoPOymZ5cBWEI3dlCLExilYRA3eETQQAbVd4wyI+roq0MpjEXgmI/oNiNxEx9uamlePm/WNIS0/2ZIwobkN66Z9rqYHGkTnxPIK0pugFAnCQ2erzdWRLlDWih0pyV9REErbkYUN8p8ieAsYf6k3yFFMVuhB3U4kdwlF8Mhhg7YNtalGGXcmOxAgOiv0ADC5gj0ACf2oAAV8hHvMPUYLCEegAcAtAAft5GiM2VA5npsNqqWjrRewwwg

cIdHHDIApwkXq6MuH8Drhh4k8eeKvH3CHxz4/0WYMDF71gxsXewQx1+ECCoxFZVLvtHmCeDCa3g5+t6jd67s5x0IgMMEIWpwjG+diBmokG0BRDohsQ1mk1wOqpp1aAIPalWNSGl50h2rDPg2NMyi10+JIqkTLXA4EMzW3ebtLN0qGDinIb1ZDjoVqErcFx2HJ1k0JdZfBFgOtfcM32FEri0sa48URck3HI1hh53WURYMH7zBUAVZZcDiFQDvdAA+

saABGTUAC/iuvxfGfg3xGw80XQMLa7DbRv4+0Xp357mdBelnM4dwLdF8CHONkuyQ5OcnuTPJ8Euni8KQnWDLexZL4Z63QmRj/h2Ew8MCNBHjUkxb9W8KRLExldRaWYmiVKyRHIimJD7dEcwVLHli8RnXHiYSMT71iy0Qkg1i2NEntik4tI3PqeUVrEMZJRfKoexJwaa18KK3HSVIiMLTiDafIj0Me10lLiuh3rXoUBQFymSe+0ooRpZNEZMdPRAd

Ixusy8n08fJZooGuCM/GhSdhxbIKQzj/HPSjh4UjgZFOAmUkYpVw8oBdKumz1guvbAMQOyDGZSQxWNb4RyAjHlBMJLLGMcwHmDxj9QiY3wcmMK68BuwaYnolRPCEOQmwNXfMaiLoJrFcUbUnER1I1YEiv2AtH9vxL6n6stieQsvgUIenWUxp3Y/PtJMGQDjBUrItrmX3mnEU6hHoOcY0MdaaSYwqwraWwzGEcN9uhk6buuJMkncpRO4xKBdzlEHi

iBAdQAMamgATocEe1QFRLgAR6AAPt3UaABnRQR6r0AATQj0ADR1oADvdG6e8Tum5s/JWwz6d+LenPEPp9Atgd9JOG/SG2IEltoDIkCHiTZZsi2dbPtlOzXZHs1Kc8NN4ZTouWUzqqGLPrwynBGEgqYVH2g8A8JT9GGPl2xn+CJSJkSmifgzELBapLyWiZEOD4h8mp4fFieMRUwzQ5oyQ2Ph+3pkJ9GZxIlYqSMEmsyixUcSkSN2NbiTTWefKSZ5g

FlzchZTkRYOyLDJV95Zy0qMg9JnH9xHIRHCwQZN+K9jwq/QjWTliOnazCse4vWQIJ4CoBAAB2qAAFbUAADFjgM9m7zGeUAagT7IekWj/ZgUrnsHI7oASrREc84U9FAkOdn578r+WDKN4ITIZWc94TnOYp5y4ZjgxLsXIvjFRipmM7dtXMgC3gxSZEiLATMzHUSW59U2VoxPJkjFKZqpamRWIHlcSh5XUhmcoUFq/sWZAHYSePKGnZ8IOPM0oRNMZ

H9i156hPliOKUkcjREFEtSVh2BprJGGTYRYJ0QmAJgT5l6M+bDXI5+sTJVHJMAYTAojDhRColUQHUABMaYACNrO/oAFDFEnoAAvVT2SaPWH3T9iwCkOVaNAV2jW6/4sOYBOgXRS4F50mxQ4ucVuLkFTw8wZeihnZyYZ1vJWSPgRnX18F0sHgOjKbKlSsZ5U7gBQvxmNy+QzcyVlVxJl5johBYuIdPPHKtSsRNMyseq2rHDzaxfC5mfXnQZCLkcIi

wodzK7ESKexFrJkcThZEJhNCo+Cvst0nQxhpZak2WQkC256Tlxu0sUcYoDZbizuxFB2mdIEGHiZQNwBHoADAlQAEI6CPQAFt2ZPBHuo0AAjNoAHh7D+YAFPzDxd7JZ5NB/J5QAOWApCl+LHRFnZ0VFKjl2cHOhy45ecquU3LHlLyuJRDLSWWC3h0aYFikpyk29C5+UudqyxljlyCJlcsqVHibCrd65yiiJKENoUVKA+bcxqUwrHK4o/s9sRbi0vf

b40axGQ/hd0v6lszBpc86kSNM7GEMyhF8mbqvNknryvgW836i3AsJNCeRINdaeUo6Et99F6yoyZssGFmTtxlipioPw+CoA3uzjCxj/PeUfjDhPywJSZ0+kAqIpQKv6cRQBlgTygeqg1UavTkJLz6SSjBairi4Iq8piMrJftA+BEKwRxEnGUSsoXCsQhhMhEdGlzHIjKgtS5iS1PGJsLaZbS7hSPN4VMyRJAitPgNNnn5DRuAy0aUMvpHQdhVFQ6a

XJMTCSqJ00qilVOIPlrSj5TYdoasp2mkcNlV8rZZqp2Vi5dZVkyJaqO9qAA9dO0aAByA3e6AASuXUaAAkuTeWiRTRgCnxV8okDmrgpQSq1ZAqAmRz/pESgQRdNHUTrp186t1YhKHbQyUJYYguXgqxUoyPguS4hY9MJXErPeDkdMVGpoVEyjIVS+NYmuan1KqZTS9hbtVnKtLuJPBTNRiE6U5rOVU89cg3h5WFr55HYsRaWogAMiK1YytQhwkxG1r

7ycyrkTPlWm8ij581VMO2t26qq1ZxkntbfO1WetB+h41cH+HnbVBZOqABQOowAAabGtgGwFJAI8bgAANQXU/Ql1Hyx6WaoCUbrLV/y7dWEpBXuj9ZAdFjdIDY0caFAqAXjbUAE1QAhNoms9WgovXJKr1+c3BX8LvWFTL4uKp9aGprnR4eykKBuZ+qbkUr5S0aOiQxI7m0qI+sIHPCgk0IF4OudMjNR0uzXCLc15IoDgWo5lFquZJawVZIqw3SKxV

si4cVLTFlUNpViyYjU2tI1yIPgiwVSQrJ24IqDFPrfaYjUOlayGNA/Jjv0FQCAA5jMACWaQAHbvubWwACN+gACBVjVi6rxcusrp+y/F6696X8ogUhKoFtq3dfav3XXDGtrWjrV92619ajNCKz1cio+HZSfVaEjJfbys0lzCQ/QYNfkpIWFKn4TmympGsolfqY1CwaVgwp83vYANCGoDWWOaUcLwNXCyDeFrHl9KotTYikRyri0ob+VaGpLSMoL6p

aq14qhSaLNHFa1pVJWveXrTr58jOiS0pVfpOo3Cr1ZdG2rRZIfmDqD1NijJu40ACS3oADi5QAIDGgAZbkxNaw/+e+N9lPTRtMm8bZuvk1Tad1MCpttHMdUSALp5O6nfTrhWoKNt6CrbZgoZZorfV+2iAEjJS5HbL4j6kNX4LIXcAtFEaj9bdrc3fq41ZMl7V3OTUljgNaaiDd1z4mwbq8PS/NSDodLxbnSiWySUrRh3MiXqZ7fDRPmlVLTG19DeV

f3B7KbTSglG8rbjrhr46NV9GonVYpU2AAyv0AD65lT0ABoygj0ABGxoAGkjW5RwEACd2oAHnExncoQk2mqvxHOoORNrM5Oi62wKvdQLrBUGyk9qejPdnoL3i60pmckzV6rM04LUcCupXXjVZAJBbN6u0he/TIKpiSV1C/XfdoDDUrGFxutEYBtVKsFCUX2lldpnaXsqulturlfUt6nIa+VRQgVa7smmirYdsikWfIpmXKS5lIe/3ejqD08Bd52Ot

ZZ2rVXdro9hO3ZQOv2XXCEgqAQAHtq6Td7oAHMjFOv1vE2DbJNvigKa9N+Vc7Jt1e4XrNtCwOqHOAB4A2AYgPrb0pXe6Xd6tQnEc/VmSw7QQuYAJBTtPg87VHh7CT631VC0pe1HKUeb3Qv6o3QpSTXL6MR5u5lfiLC3b6bd9Su3dyti2O6wdx+iHafqkVTSPd1azee9UR0LS5lu8h/RpLjKDw2cYehjhVr2lZlP9kou2j/tOnhsh1AdQAHdu9448

YAFnEzxgQOKbqMp1RejECXtZ3Sb4DFq3nlup52Ka69oKsw5YZsN2HimTh3A53qsGmbbBsM3Kf3oDWEgEgaus7c+vaLvAdd0+spe5oZqG6alvm7uWbo+0gb88nE77ayq33W7ItcGwRfbp33iGj9gyyHXzJXl86B868wgopJv2KLZUnRR4p2UPmiIew9xN4C/u0PEddDXarvu2E7BmKaOvfXcXHoOUGzAARdFACEegARbtTZDhjgFOvPGAAyWOcN/y

AFMB1dS9JtEIG5NSBwFTXrtVoH5t5QQ8csbWMbHtjZ4vY2EYsGbaFg223OdEfRW3rkug+ioG8BH1JH7NmuvkGYuc2kqWoNUzIwAy83tzauncpfW9oSGTFo+6+/g79sEMVHd98GtYgftB11GXdS8t3bIfGUvUey3uiWQmoD1o71DuHfoG8BWXbSqN7+mjeqsMP5lY9Oqpjm8FQCAAlfRfnvdAAt/HH9IDTOw46XpAUeHZNXh7ncgZdF+HlNAgvk4K

ZFNim3jiSqXZ8Zl1jsYjGK/1WQd+RUHCJVci7WCjxnXbqp5K79XGAakL7ODr2lhTwYKMW6ftVuokQSeEN76ENXpqWpzOd0n6STZ+5oyTjkltgqTD5NsAsrUU4cMdTYCla/o7WiiP9ExzWUYf7UmHht1wi6YAFP3QAMoJ8woOoACuVQAKbm+x1w0AuOP+KZTnO841XsuMoHQz6BswwWaLNln29Gc949qb5BfGsFPx+XQadIP/G76bwRI9QeSO7tzC

lphg1KXSPMHYTlS0mTkcX0Uzq0729qXwc6lYnyjAOyo3mtEMO7/TTu26hN3GlQ7+ZoZiZdGEjOToPisqkjYHrkR1Q21LJ8PWybx20av9GZtGlmc6YLGA6gARbdAA5cYI8XJgACVNAAU9GABa0xz2ABZ5UAARmYAFvUwAPXOFZ6A1KfZ21mK9iBhszaquOoHx8LZgCyBbAtQXYLiF1C52fdURckVOpwg9eos1FyXE4Ab6D6jgBwA8QIUbgNaGgA/g

Mg+nV7MMAYCEAEAFAYamBo33uwUQKCH4PDsAmDV/IMobMPoDxDJIEhCAeYJpYIlWjFLWQZS+kAktwNMTHpk4XpagAGX9APwGDTiYUt+olLKltS5PPUAABpBAE+AgAZBHAQgfQD0F0v2X9Ljlq2MJt9RWBNAWIdpNNqAIjz/L4OSy05ZnkO7YrDl9IKvTbEYxkrgV9IDWSDMFBMrFllS5JwwsSlhLA1AKwVfSBFXyBRwDDnZbisqW3khwznqVfMvx

WySTVRtvlcstwhGqFFZqiQTMvlXLLLuecICiWh+Wyr9VyqznDSsshx0FQbABcGxDcbuAzgJIOYXGBfA2hcYB2MbTyvMBFr9wfAKuxExvB6J4oGPHGHBRXtwaeVowPxt8uBlvQBAIQING0B1QiVc4rjF1ZUtpXy+YqOafaQw3NtiAKnN+MJdBAkA8QCAcZiVbyuQ3iA8neSD1dwCaBggRZYGyQBSQFRag+AIqKQGUCAgAAFO2A1S8BVu+sNuMcAsw

ABKNEOvWUA6hfUrQQm7gBJvH1eAq1E2PyGpsJA6b31ya/5ASsIAaydZXTH1hFXr0DQzbM3AVEyCo30b3Ac4K9ZrZEBYbqAZW+fVbYCW0AmtgUHxTN5K3SAKt04PoF9TXBSAoYJhEbZNurAzbpAC2yjbRuGRNb31uwAACs/EzAHEK2zgBI2EATtxW+wx9TtBCAjAecPxqU5PX8VhINIKHfRg1shApIAwKNYESZnidl6NUAYBxBx2xbgY4DDBFDvh3

I7/0b644GYAK3rYxbYWLJ0yBCBGSIMCApBkCDogmAmQXGMtkrtA20w+JQOy7eNt2pQ9snEgLfh9u4AOLqGPu5JD1t5WiwmAHO8EHjscB/b3yyKbO0bjQbggVoB0CAAdBAA==
```
%%