ผลต่างระหว่างรุ่นของ "Probstat/week14 practice 1"

จาก Theory Wiki
ไปยังการนำทาง ไปยังการค้นหา
 
(ไม่แสดง 2 รุ่นระหว่างกลางโดยผู้ใช้คนเดียวกัน)
แถว 1: แถว 1:
 +
: ''This is part of [[probstat]]''.
 +
 
== Least square estimators ==
 
== Least square estimators ==
 
1. You should work in pairs.  Try the following for <math>n=3, 5, 20, 100</math>
 
1. You should work in pairs.  Try the following for <math>n=3, 5, 20, 100</math>
แถว 10: แถว 12:
 
* Give the data to your partner.
 
* Give the data to your partner.
  
* After you get the data from your friend, find the least square estimators for <math>\alpha</math> and <math>\beta</math>.
+
* After you get the data from your friend, find the least square estimators ''A'' and ''B'' for <math>\alpha</math> and <math>\beta</math>.
  
 
* Ask for the correct parameters and see how close you get to those parameters as we increase the size of the samples <math>n</math>.
 
* Ask for the correct parameters and see how close you get to those parameters as we increase the size of the samples <math>n</math>.
  
 
== Multiple linear regression ==
 
== Multiple linear regression ==
 +
Follow the same procedure as in the previous section, but try with 2 variables <math>x_1</math> and <math>x_2</math>.  I.e., you pick parameter <math>\alpha, \beta_1, \beta_2</math> and generate samples from the following formula
 +
 +
<center><math>Y = \alpha + \beta_1 x_1 + \beta_2 x_2 + e</math>.</center>

รุ่นแก้ไขปัจจุบันเมื่อ 01:14, 27 พฤศจิกายน 2557

This is part of probstat.

Least square estimators

1. You should work in pairs. Try the following for

  • First, you pick a parameter and and generate a sample of size using the formula:

Use random values for in the range that you like. For the error , you can generate a random error based on a normal distributions.

  • Give the data to your partner.
  • After you get the data from your friend, find the least square estimators A and B for and .
  • Ask for the correct parameters and see how close you get to those parameters as we increase the size of the samples .

Multiple linear regression

Follow the same procedure as in the previous section, but try with 2 variables and . I.e., you pick parameter and generate samples from the following formula

.