原创 attention機制介紹，易懂
以下文字轉自網址： https://towardsdatascience.com/attn-illustrated-attention-5ec4ad276ee3#0458 Attn: Illustrated Attention At
2020-06-17 06:40:27
7
原创 Understanding LSTM and its diagrams
the Pic is from the following site: https://medium.com/mlreview/understanding-lstm-and-its-diagrams-37e2f46f1714
2020-06-17 06:40:27
原创 理解一致性
最近看到知乎文章關於“無偏性”和“一致性 ”的解釋，覺得寫得很好。轉載一下。- 叄晟的文章 - 知乎 https://zhuanlan.zhihu.com/p/66658725 無偏性：這個定義設定的目的在於判斷統計量在每次試驗中的取值對
2020-06-17 06:40:27
5
原创 Typing bold symbols in math blocks in Typora|(在typora中輸入加黑數學符號）
In Typora, sometimes u would like to type bold math letters, which are symbols of a vector or a matrix. You can follow
2020-06-17 04:50:06
原创 Difference between accuracy and precision
accuracy：準確 precision：精確 Accuracy and precision are two important factors to consider when taking data measurements.
2020-06-17 04:50:06
原创 將字符型分類變量數據轉換爲數字格式
假設我們有如下dataframe，名字爲student_weight： ordered_height = ['Short', 'Medium', 'Tall']#定義順序，這裏是這個次序可以對應，0：short；1,：Medium；2：
2020-06-17 04:50:06
原创 Difference Between Labor and Human Capital
When talking about production function, you should have clear delimitation of "labour" and "human capital". The follow
2020-06-17 04:50:06
原创 工具變量法（instrumental variable method）
傳統ols中的hypothesis要求xix_ixi與uiu_iui無關。但是在實際中很難滿足這個假設，有時候因變量（在單方程模型中就是內生變量）也會反過來影響自變量。ols估計將是有偏和不一致的，（有偏指的是參數估計值於期望
2020-06-17 04:50:06
原创 關於keras 中輸出模型網絡圖的說明-how to plot the lstm model graph and save it to a file
在keras中，可以通過Model visualization模塊進行作圖。以下是來自keras文檔的說明。 Model visualization Keras provides utility functions to plot a K
2020-06-17 04:50:06
原创 lstm模型的5個步驟
the 5 steps in the LSTM model life-cycle in Keras that we are going to look at. Define Network Compile Network Fit Netw
2020-06-17 04:50:06
原创 Multiple Input and Output Models in keras
Multiple Input and Output Models The functional API can also be used to develop more complex models with multiple input
2020-06-17 04:50:06
原创 Introduction of moving block bootstrap （MBB）algorithm
Because we can not use usual bootstrap sampling method to get subsamples from time series dataset, then the MBB was
2020-06-17 04:50:06