DATA EXPLORATION

Within this data set there are 2276 observations of 16 variables. The main focal point of this data is that we want to predict the target wins that a team will have over a given parameters.

To first attack the data set there was some basic cleaning to remove the unnecessary naming within the columns. We then did some exploration summary of each column as well as the missing values within each. (ALL ACTIONS DONE TO TRAINING SETS DONE TO TESTING).

DATA PREPARATION

In the preparation of the data for analysis I utilized various techniques to best decide how I would proceed with my analysis. One of these methods included creating a correlation heat map that would be essential in allowing me to better understand the data. This is significant because it would allow me to make a more informed decision as to what type of model I would create on the data set.

In continuation of preparation I utilized plots of all the variables against the target variable to see any specific linear relationships between them. Overall there weren’t very much direct linear relationships.

BUILD MODELS

To build the build the models I will go with three approaches. The first will be a basic approach that will give us a model that is not tampered with. The second model chosen was normalized as well as scaled. This would be able to give us a model that had the stronger assumptions of regressions. The final model would be one that incorporated backward propogation. This would be one that removed variables one at a time with p values > 0.05.

SELECT MODELS

The model selection here we will go with will be the seceond model. In terms of selecting a model we will always go for the best performance because that should give us the best results in real world scenarios. We don’t want to be bias in our decision as it could hinder us going further. What we notice in the model however is that we didnt have a perfectly normal dataset through the residual plots. The Q-Q plot also showed various skewedness through the tailends. Overall this was similar throughout the models so through chosing the strong R-squared value we selected the model with the strongest predictor of future variables.

APPENDIX

library(tidyverse)

## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr     1.1.3     ✔ readr     2.1.4
## ✔ forcats   1.0.0     ✔ stringr   1.5.0
## ✔ ggplot2   3.4.3     ✔ tibble    3.2.1
## ✔ lubridate 1.9.2     ✔ tidyr     1.3.0
## ✔ purrr     1.0.2     
## ── 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(dplyr)
library(naniar)
library(mice)

## 
## Attaching package: 'mice'
## 
## The following object is masked from 'package:stats':
## 
##     filter
## 
## The following objects are masked from 'package:base':
## 
##     cbind, rbind

library(corrplot)

## corrplot 0.92 loaded

library(ggplot2)
library(tidyr)
library(Metrics)
library(ggfortify)
set.seed(1234)

Load Training and Testing data

trainraw <- read_csv('https://raw.githubusercontent.com/kglan/MSDS/main/DATA621/Assignment1/moneyball-training-data.csv', col_names = TRUE)[,-1]

## Rows: 2276 Columns: 17
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## dbl (17): INDEX, TARGET_WINS, TEAM_BATTING_H, TEAM_BATTING_2B, TEAM_BATTING_...
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

testraw <- read_csv('https://raw.githubusercontent.com/kglan/MSDS/main/DATA621/Assignment1/moneyball-evaluation-data.csv',col_names=TRUE)[,-1]

## Rows: 259 Columns: 16
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## dbl (16): INDEX, TEAM_BATTING_H, TEAM_BATTING_2B, TEAM_BATTING_3B, TEAM_BATT...
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.