soap945/infs692

Data Science with R

INFS692

Data Science with R

There is a problem initialize the h2o on chunck 13 on Dec 16th, which does not allow me to knit the pdf. So, Model1 report is based on the first version which did not convert the Institution to categorical variables. Model1 markdown file is updates which has the categorical value conversion set

The final submission is in the final project folder

There are three models:

1. Model1 with stacking
2. Model2 with DNN
3. Model3 with K-means, hierarchy and model-based

Model 1 libraries:

library(tidymodels)
library(plyr)
library(caret)
library(caretEnsemble)
library(readr)
library(ggplot2)
library(tidyverse)
library(rpart) # for fitting decision trees
library(ipred) # for fitting bagged decision trees
library(pROC)
library(h2o)
library(ggcorrplot)
library(ROCR)
library(recipes)
remotes::install_github("kforthman/caretStack")

Model 2 libraries:

library(tensorflow)
library(keras)
library(caret)

Model 3 libraries:

library(dplyr)
library(factoextra)
library(cluster)
library(mclust)

There are alternative visualizations and approaches written in the code file
Uncomment the cells to show the results
Some printed are too long, so delete the invisible() function to check the whole results