iT邦幫忙

2019 iT 邦幫忙鐵人賽

DAY 8
0
自我挑戰組

機器學習與數學天天玩系列 第 8

Day 8-機器學習與數學天天玩-PCA-Statistical Introduction: Linear Transformation Part 2

The brief structure leading to the milestone of PCA is as below:

  1. Statistical Introduction
  2. Transformation of Vectors in Spaces
  3. Orthogonal Projectio

We talked linear transformation of shift yesterday, let's look into the scale today.

Assume the original dataset and its mean value is as follows:
income_data = [23000, 50000, 40000]
E[income_data] = (23000+50000+40000)/3 = 37666.67

Scale: Because of the economic miracle, the government decides to have every citizen's income double.
*income_scale_data = [46000, 1000000, 80000]
E[income_scale_data] = (46000+100000+80000)/3
= 75333.33 = 37666.67 * 2 = E[income_data]2

By observing the change of the value, we got a relationship between the orginal dataset and the scaled dataset, that is, E[income_scale_data] = E[income_data]*2.

Having the formular be more general, it as follows:
E[D*b] = E[D]*b, where b is the number of scaled value


上一篇
Day 7-機器學習與數學天天玩-PCA-Statistical Introduction: Linear Transformation
下一篇
Day 9-機器學習與數學天天玩-PCA-Statistical Introduction: Linear Transformation Final Summary
系列文
機器學習與數學天天玩13

尚未有邦友留言

立即登入留言