Supervised Learning Algorithms $($SVM$),$ YOU NEED TO WRITE A PYTHON CODE PLEASE READ THE INSTRUCTIONS CAREFULLY, AND PLEASE DO NOT SUBMIT AN ANSWER IF IT'S INCOMPLETE $.$FIRST YOU NEED TO IMPORT THESE LIBRARIES

import pandas as pd

from sklearn.model$\_$selection import train$\_$test$\_$split from sklearn.metrics import accuracy$\_$score, classification report From sklearn.metrics import roc$\_$auc$\_$score from sklearn.preprocessing import StandardScaler from sklearn.neighbors import KNeighborsClassifier from sklearn.tree import DecisionTreeClassifier from sklearn.ensemble import RandomForestClassifier from sklearn. linear$\_$model import LogisticRegression

THEN YOU NEED TO READ THE DATASET USING : df$=$ pd$.$read$\_$csv$("$Dataset#$4.$ csv$")$

df$.$drop$\_$ duplicates$($inplace $=$True$)$

YOU NEED TO TAKE A LOOK ON THE FIRST TABLE AND MAKE TWO COPIES $($WE NEED TWO TABLES ONE FOR EACH PHASE $,$ BEFORE OVERSAMPLING AND ONE AFTER OVERSAMPLING$)$

FOR THE FIRST PHASE TAKE THIS CODE AND JUST ADD LOGISTIC REGRESSION TO IT

import pandas as pd

from sklearn.model$\_$selection import train$\_$test$\_$split

from sklearn.metrics import accuracy$\_$score

from sklearn.neighbors import KNeighborsClassifier

from sklearn.tree import DecisionTreeClassifier

from sklearn.ensemble import RandomForestClassifier

df $=$ pd$.$read$\_$csv$(\text{'}$Train$1.$csv$\text{'})$

mean$=$ df$["$X$\_12"].$mean$()$

df$.$fillna$(\{"$X$\_12"$:mean$\},$inplace$=$True$)$

drop$\_$cols $=[\text{'}$INCIDENT$\_$ID$\text{'},$'DATE'$]$

df$.$drop$($columns$=$drop$\_$cols, inplace$=$True$)$

X$\_$train, X$\_$test, y$\_$train, y$\_$test $=\backslash$

train$\_$test$\_$split$($df$.$drop$(\text{'}$MALICIOUS$\_$OFFENSE', axis$=$'columns'$),$ df$[\text{'}$MALICIOUS$\_$OFFENSE'$],$ train$\_$size$=0\mathrm{.}80,$ random$\_$state$=2)$

############################# KNN #########################################

model $=$ KNeighborsClassifier$($n$\_$neighbors$=1)$

model.fit$($X$\_$train, y$\_$train$)$

pred$\_$test$\_$knn$=$ model.predict$($X$\_$test$)$

pred$\_$train$\_$knn$=$ model.predict$($X$\_$train$)$

print$("$K$-$Nearest Neighbors accuracy score$($test$)$ : $",$accuracy$\_$score$($y$\_$test, pred$\_$test$\_$knn$))$

print$("$K$-$Nearest Neighbors accuracy score$($train$)$ : $",$accuracy$\_$score$($y$\_$train, pred$\_$train$\_$knn$))$

print$()$

######################### Decision Tree ################################

tree$\_$clf $=$ DecisionTreeClassifier$($max$\_$depth$=16)$

tree$\_$clf$.$fit$($X$\_$train, y$\_$train$)$

pred$\_$test$\_$tree$=$ tree$\_$clf$.$predict$($X$\_$test$)$

pred$\_$train$\_$tree$=$ tree$\_$clf$.$predict$($X$\_$train$)$

print$("$Decision Tree accuracy score$($test$)$ : $",$accuracy$\_$score$($y$\_$test, pred$\_$test$\_$tree$))$

print$("$Decision Tree accuracy score$($train$)$ : $",$accuracy$\_$score$($y$\_$train, pred$\_$train$\_$tree$))$

print$()$

######################### Random Forest ################################

rnd$\_$clf$=$ RandomForestClassifier$($n$\_$estimators$=120,$ max$\_$leaf$\_$nodes$=200,$ n$\_$jobs$=-1)$

rnd$\_$clf$.$fit$($X$\_$train, y$\_$train$)$

pred$\_$test$\_$rf $=$ rnd$\_$clf$.$predict$($X$\_$test$)$

pred$\_$train$\_$rf $=$ rnd$\_$clf$.$predict$($X$\_$train$)$

print$("$Random Forest accuracy score$($test$)$ : $",$accuracy$\_$score$($y$\_$test, pred$\_$test$\_$rf$))$

print$("$Random Forest accuracy score$($train$)$ : $",$accuracy$\_$score$($y$\_$train, pred$\_$train$\_$rf$))$

NOTE TAKE THE NUMBERS OF TRAINING AND TESTING FROM THE FIRST SCREENSHOT

AND WE NEED TO DO STANDARD SCALING ON X TRAIN AND X TEST $($X TRAIN USE FIT.TRANSFORM, X TEST USE TRANSFORM ONLY$)$

AND THEN WE NEED TO PRINT THE RESULTS $1.$ JUST PRINT THE ACCURACY OF THE TRAINING AND TESTING AND PRINT THE CLASSIFICATION REPORT THE CLASSIFICATION REPORT CONTAINS $($Precision$,$Recall,F$-$Score$)$ FOR EACH CLASS IN THE LABEL $)$ this is before oversampling $($phase$1)$

$***$For after oversampling

U can use this code

from imblearn.over$\_$sampling import SMOTE

import pandas as pd

train$\_$df $=$ pd$.$read$\_$csv$("$train$.$csv$")$

train$\_$Y $=$ train$\_$df$[\text{'}$attack$\_$category'$]$

train$\_$x $=$ train$\_$df$.$drop$([\text{'}$attack$\_$category','attack$\_$type', 'protocol$\_$type','service','flag'$],$ axis$=1)$

print$($train$\_$Y$.$value$\_$counts$())$

sm $=$ SMOTE$($sampling$\_$strategy$=$'auto', random$\_$state$=0)$

train$\_$x$\_$sm$,$ train$\_$Y$\_$sm $=$ sm$.$fit$\_$resample$($train$\_$x$,$ train$\_$Y$)$

print$($train$\_$Y$\_$sm$.$value$\_$counts$())$

You just need to do scaling and print the results

For the AUC in the tables we use this print statement to print it

y$\_$pred$-$DT$.$ predict$\_$proba$($x$\_$test$)$

print $($roc$\_$auc$\_$score$($y$\_$train, y$\_$pred, multi$\_$class$-$'ovr'$))$

Here is the statement for the classification report use it : classification$\_$report $$y$\_$test, pred$\_$test$\_$tree$$ You should submit the following:

$1-$ Python code file

$2-$ Documentation with the description of the steps you followed and an explanation of the results you got.

The dataset description

The Dataset.csv file is a clean dataset of $253,680$ survey responses to the CDC$\text{'}$s BRFSS$2015.$ The target variable Diabetes$\_012$ has $3$ classes. $0$ is for no diabetes or only during pregnancy, $1$ is for prediabetes, and $2$ is for diabetes.

This dataset has $21$ features and $253680$ records. The following is the description of the features:

$\backslash$table$[[$Variable Name,Type,Description$],[$Diabetes$\_$binary,$$ Binary,$0=$ no diabetes $1=$ prediabetes or diabetes$],[$HighBP$,$Binary,$0=$ no high BP$1=$ high BP$],[$HighChol$,$Binary,$0=$ no high cholesterol $1=$ high cholesterol$],[$CholCheck$,$ Binary $,,\backslash$table$[[0=$ no cholesterol check in $5$ years $1=$ yes cholesterol check in $5],[$years$]]],[$BMI$,|$Integer$|,$Body Mass Index$],[$Smoker$,$ Binary $,,\backslash$table$[[$Have you smoked at