Application_On_GlassData.R

# Data set from https://machinelearningmastery.com/machine-learning-datasets-in-r/
library(pracma)
library(mlbench)
# Glass Identification Database
data(Glass)
dim(Glass)
levels(Glass$Type)
head(Glass)

# Put Glass data into dataframe
df <- data.frame(Glass)

# Remove NA values by removing row
df <- na.omit(df)

# init = rep(.1, ncol(df))

model1 <- glm(Type ~ RI + Na + Mg + Al + Si + K + Ca + Ba + Fe, data = df, family = binomial)

init = model1$coefficients + .05*rnorm(ncol(df), 0, 1)

df <- cbind(free = 1, df)

# y <- rbinom(nrow(df), 1, p.true) 

# df[ncol(df)] <- y
# Turn df into a matrix to make compatible with algorithm

print(unique(df[c("Type")]))

df <- data.matrix(df)

# df_norm <- as.data.frame(lapply(df[:], min_max_norm))
# df_norm <- scale(df)

# print(df)


# # Apply the stochastic newton algo
# print(tail(stochastic_newton_algo(df,init))) # Output gives lots on NaNs (could be dividing by 0?)
# 
# # Apply the truncated stochastic netwon algo
# print(tail(trunc_stochastic_newton_algo(df, 10^(-10), .49, init)))
# 
# model1 <- glm(Type ~  RI + Na + Mg + Al + Si + K + Ca + Ba + Fe, data = Glass, family = binomial)
print(model1$coefficients)

print(sgd2(df))

print(tail(df))