Homework 7

library(tidyverse)

## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr     1.1.4     ✔ readr     2.1.5
## ✔ forcats   1.0.0     ✔ stringr   1.5.1
## ✔ ggplot2   3.5.2     ✔ tibble    3.3.0
## ✔ lubridate 1.9.4     ✔ tidyr     1.3.1
## ✔ purrr     1.1.0     
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag()    masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors

library(readxl)
library(pastecs)

## 
## Attaching package: 'pastecs'
## 
## The following objects are masked from 'package:dplyr':
## 
##     first, last
## 
## The following object is masked from 'package:tidyr':
## 
##     extract

library(lmtest)

## Loading required package: zoo
## 
## Attaching package: 'zoo'
## 
## The following objects are masked from 'package:base':
## 
##     as.Date, as.Date.numeric

library(MASS)

## 
## Attaching package: 'MASS'
## 
## The following object is masked from 'package:dplyr':
## 
##     select

setwd("C:/Users/KaeRo/Desktop/R Studio/Reseach Data Selection")
library(readxl)
district <- read_excel("district.xls")
Cleaned_district<-district %>% drop_na()

model1<-lm(DPSTTOSA~DPSAMIFP+DPETWHIP+DPETALLC+DZRVLOCP, data=Cleaned_district)

Check the following assumptions: a) Linearity (plot and raintest)

plot(model1,which=1)

raintest(model1)

## 
##  Rainbow test
## 
## data:  model1
## Rain = 0.87665, df1 = 162, df2 = 156, p-value = 0.7966

Does it meet Linearity: Somewhat ? The line itself isnt the most straight but has a good portion of linearity, and the p value in the rainbow test is high

b) Independence of errors (durbin-watson)
d) Normality of residuals (QQ plot, shapiro test)
r plot(model1,which=2)
r shapiro.test(model1$residuals)
## ## Shapiro-Wilk normality test ## ## data: model1$residuals ## W = 0.99459, p-value = 0.3117 Does it meet Normality of residuals: A lot of the dots are on the line, and the p value is above .05, so the residuals are normal

5. No multicolinarity (VIF, cor)

vif(model1)

## DPSAMIFP DPETWHIP DPETALLC DZRVLOCP 
## 5.261047 5.237581 1.285218 1.149458

None of the VIF’s are over 10 (and only 2 are over 5), so the variables are not related to each other

So to review: a) The model is somewhat linear b) The errors are somewhat independent c) We can’t fully reject that the model is not homoscedastic d) The residuals are normal e) The variables are not related to each other

I don’t think we can fully reject any of the assumptions, the only one that is a bit iffy is Homoscedasticy, which we can do a log on the model to mitigate

okay I have been trying to do the log of this data set but there is something that has infinite values and i have tried to so many times to transform the data and I keep getting the error message (orginally looked like model1_log<-lm(log(DPSTTOSA) ~ log(DPSAMIFP) + log(DPETWHIP) + log(DPETALLC) + log(DZRVLOCP)), Error in lm.fit(x, y, offset = offset, singular.ok = singular.ok, …) : NA/NaN/Inf in ‘x’

I have googled the problem, I’ve rewritten the code to take out any NAs, Infinites, or non numbers, so honestly I’m not quite sure what to do from here, but I do know that taking the log of the data will make it more homoscedastic.

```