forecasting:meeting_minutes_august_31_2016

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

forecasting:meeting_minutes_august_31_2016 [2016/09/08 03:31]
jobatake created
forecasting:meeting_minutes_august_31_2016 [2021/09/19 21:59]
Line 1: Line 1:
-Went over Masaki’s use of 3D plots for zenith-angle normalized plots, etc. 
-He used his own parsing for the date 
-Use of np.meshgrid() and Axes3D() 
-New explanation of basics of linear regression/​machine learning. 
-W is the slope 
-b is the intercept 
-X [n x m] 
-2D matrix of inputs (row = sample; column = data) 
-W [n x 1] 
-1D column of weights 
-D [m x 1] 
-1d column of stuff 
-Y = XTW + b*1 (where 1 is a 1 column vector) 
-E = D - Y 
-Error 
-J(W, b) 
-= (1/2m) * ETE 
-= (1/2m) * (|D|2 - DTY - YTD + YTY) 
-= (1/2m) * (|D|2) - (1/m) * DT * (XTW + b*1) + (1/2m)(XWT + b*1T)(XTW + b*1) 
-dJ/dW = 0 and dJ/db = 0 → W = C-1Q 
-dJ/db 
-= (-1/m)DT1 + (1/​2m)1T(XTW + l1) + (  
- 
-Build functions 
-Explanation of math 
-Handout with math and functions and documentation. 
-Functional specification 
-Focus on “legacy” and the base for continuation. 
-Go by “black box” before understanding. 
-PDFs or slides for presentation 
-1. We want reference for code and the supporting concepts. 
-2. We want to focus on learning specific algorithms. 
-3. We want to have material to have a sense of continuation and legacy in a project and also to support understanding and learning. 
-Setup for continuity and focus on code. 
- 
-Main Future Points 
-Git tutorial? 
-Code + Concept Documentation 
-We want to learn Python in-code documentation 
-Meeting Minutes 
- 
-Code + Concept Documentation Skeleton/​Template:​ 
-Title 
-Function Title 
-parameters/​types 
-What it does 
-Explanation 
-Primary Overview 
-Math explanation 
-Notes/​Remarks 
-(optional) 
- 
  
  • forecasting/meeting_minutes_august_31_2016.txt
  • Last modified: 2021/09/19 21:59
  • (external edit)