Stat 306
Midterm exam
February 2, 2010.
Instructions: Please write your name and student number on the front page of the exam booklet.
A two-sided sheet of notes and a non-programmable calculator may be used. The 4 problems may
be done in any order.
(25)
Stat 306 (Fall 2010)
Homework 1.
For questions 1 and 2, use R (or Excel or another software) to get the least squares equations and SD
of residuals, and summary statistics (sample means, SDs, correlation). Then other calculations can be done
using a calcu
Question 1.
The specic activity of the enzyme sucrase was measured by extracting a portion of the intestines of 24
patients who underwent an intestinal bypass. After the sections were extracted they were homogenized
and analyzed for enzyme activity. Two d
THE UNIVERSITY OF BRITISH COLUMBIA
FINAL EXAMINATION, April 2010
STAT 306, Finding Relationships in Data
TIME: 2 1/2 hours
THIS EXAMINATION CONSISTS OF 4 PAGES. THERE ARE 7 PROBLEMS FOR A TOTAL OF 100
POINTS. PLEASE CHECK TO ENSURE THAT THIS PAPER IS COMP
UNIVERSITY OF BRITISH COLUMBIA
STAT 306 : Quiz 1
Date : 08 Oct 2009
Quiz 1
1. For simple linear regression models fitted by the method of
least squares, show that
sum_cfw_i=1^n Xiri = 0
2.
In a multiple regression with two explanatory variables, the
total
Stat 306 (Fall 2010)
Homework 1.
For questions 1 and 2, use R (or Excel or another software) to get the least squares equations and SD
of residuals, and summary statistics (sample means, SDs, correlation). Then other calculations can be done
using a calcu
STAT 306 Quiz 2
1.
A small experiment was conducted to fit a multiple regression equation relating the
yield y to temperature x1, reaction time x2 and presence/absence of catalyst. The results,
with temperature and reaction time coded for simplicity, were
UNIVERSITY OF BRITISH COLUMBIA
STAT 306 : Quiz 3
Date : 19 Nov 2009
1a.
Before performing a multiple linear regression fit to regress the
response variable on the explanatory variables, we usually perform the R
command pairs()to visualize the correlation
Stat 306 (Fall 2010)
Homework 3.
This homework is mainly on using R for nding best regression equations of dierent subset sizes of
explanatory variables, and comparing them with adjusted R2 , residual SE and cross-validated root mean
square (prediction) e
Stat 306 (Fall 2010)
Homework 4.
Plots, logistic regression and principal component analysis
Annual nancial data are collected for rms approximately 2 years prior to bankruptcy and for nancially
sound rms at about the same point in time. There are four ex
Stat 306 (Fall 2010)
Homework 2.
1. Multiple regression; see salary.dat.txt (n = 150) at course web site
y=salary in 1000s
potential explanatory variables are:
age = age
experien = #yrs experience
yrsassoc = #yrs at associate level
yrsfull = #yrs at full
Outlines of solutions for math questions
2. R2 is the sample squared correlation of yi and yi .
(a1) n1 yi = y. It was derived in the lectures that n1 ri = n1 (yi yi ) = 0 [this was from taking
n
partial derivative of i=1 (yi 0 1 xi1 . . . p xip )2 with r
Stat 306
Midterm exam
March 30, 2010.
Instructions: Please write your name and student number on the front page of the exam booklet. A twosided sheet of notes and a non-programmable calculator may be used. The 3 problems may be done in any
order.
(16) 1.
<!DOCTYPE html PUBLIC "-/W3C/DTD HTML 4.01 Transitional/EN">
<!- saved from url=(0043)file:/C:/Users/rubys/Downloads/hdata.html ->
<html><head><meta http-equiv="Content-Type" content="text/html; charset=windows-1252"><title></title>
<meta http-equiv="Cont