ภาพหลุดด้านหน้าและหลังของกล้อง Canon EOS R3 mirrorless โดย layout ปุ่มต่างๆคล้ายกับกล้อง 1DX
//ม่วนใจ๋
「continuous learning deep learning」的推薦目錄:
- 關於continuous learning deep learning 在 Facebook 的最佳貼文
- 關於continuous learning deep learning 在 FinLab財經實驗室 Facebook 的精選貼文
- 關於continuous learning deep learning 在 Dan Lok Facebook 的精選貼文
- 關於continuous learning deep learning 在 GitHub - ContinualAI/continual-learning-papers 的評價
- 關於continuous learning deep learning 在 Lifelong ML 的評價
- 關於continuous learning deep learning 在 Is continuous learning possible with a deep convolutional ... 的評價
continuous learning deep learning 在 FinLab財經實驗室 Facebook 的精選貼文
【時間序列量化交易 - FinLab實體活動】
機器學習模型用於實際交易,不容易!
FinLab在此次的研討會中,介紹了很多小撇步,讓機器學習效果更好,更能夠應用於實際交易。以下是重點摘要:
📕【Labeling】 是很容易被大家忽略的部分,以「持有N天後報酬率」當做label真的好嗎?
我們翻閱 paper 列舉更好的方法,
Tripple barrier:
[Prado 2018] Advances in Financial Machine Learning
Continous trading signal
Continuous trading signals
[Dash 2016] A hybrid stock trading framework integrating technical analysis with machine learning techniques
Trading Point decision
[Chang 2009] Integrating a Piecewise Linear Representation Method and a Neural Network Model for Stock Trading Points Prediction
📕【 CNN神經網路】除了做出好的label,我們也介紹了CNN要如何實際用於trading:[Sezer 2018] 產生很多不同參數的技術指標,並且將這些技術指標做成圖片,來預測交易訊號,非常的有趣!
[Sezer 2018] Algorithmic Financial Trading with Deep Convolutional Neural Networks: Time Series to Image Conversion Approach
📕【 Feature importance】通常我們會利用 feature importance來篩選重要的feaature,但是我們平常真的做對了嗎?這部分我就先賣個關子XD
研討會頭影片:
https://drive.google.com/file/d/18dEAalouKvWAtlZSUjc8X9_i8uOHMwci/view?usp=sharing
假如你看到這裡,恭喜你!告訴你一個小秘密,就是我們要辦一場教學活動,由於是第一次實驗性質試辦,此活動酌收茶點費100元,設備場地感謝Fugle的贊助!
當天是個三個小時的小課程,會介紹如何用上述的內容做出一個有效的機器模型策略喔!(需要自行帶筆電)
時間:7/27號 (星期六) 下午 2 點到 5 點
地點:台北捷運西門站五號出口3分鐘路程
名額:25名(白老鼠XD)
金額:100元用來吃吃喝喝(因為是白老鼠,才是這個價格XD)
條件:完成AI股票理專課程的「第 1 章單元 2」
(課程網址:https://hahow.in/cr/finlab-ml)
報名辦法:
這個星期天(7/14)晚上10點準時收看finlab粉絲團,我們會在此時發佈表單,採先搶先贏的方式報名喔!(場地空間有限~不好意思囉)
continuous learning deep learning 在 Dan Lok Facebook 的精選貼文
One of the unique qualities all highly successful people have in common…
… is the everlasting hunger to constantly learn.
After speaking with @brianscudamore, founder of 1800gotjunk, I quickly realized that Brian was so humble regardless of his success and constantly learning.
You see, life is about continuous growth, but most people look at the most successful people and get discouraged.
They only see the top of the mountain, but they never see the path that’s right in front of them.
But what if you could read a book, take a course, listen to a podcast that could improve your income (and your life) by just 2%... 5%... 7%?
And what if you didn’t just study one book, course, or podcast…
… what if you studied 2… 3… 5?
That’s the mindset I want you to have, and I know you have it deep down within you!
Rather start small than never start.
I also highly recommend reading Brian's book, which you can get on Amazon here: bit.ly/wtf_book.
What is the one thing you want to improve in your life right now? Comment below.
continuous learning deep learning 在 Lifelong ML 的推薦與評價
One of the most significant and challenging open problems in Artificial Intelligence (AI) is the problem of Lifelong Learning. Lifelong Machine Learning ... ... <看更多>
continuous learning deep learning 在 Is continuous learning possible with a deep convolutional ... 的推薦與評價
In general, is continuous learning possible with a deep convolutional neural network, without changing its topology? ... <看更多>
continuous learning deep learning 在 GitHub - ContinualAI/continual-learning-papers 的推薦與評價
NIPS 2016 Workshop on Continual Learning and Deep Networks, 2016. Progressive Neural Networks by Andrei A Rusu, Neil C Rabinowitz, Guillaume Desjardins, Hubert ... ... <看更多>