use r studio Part $1-$ Data Exploration $(40\%)$

Descriptive Statistics $(15\%)$

$1.(2\%)$ What is the average value of the "motor$\_$UPDRS" score?

$2.(3\%)$ What is the average value of the "total$\_$UPDRS" score for subjects over the

age of $70?$

$3.(5\%)$ Calculate the following statistics for the "Jitter$(\%)"$ column:

rosting o Mean:

o Median:

o Min:

Max:

Standard Deviation:

Interquartile Range $(75$th percentile $-25$th percentile$)$:

$4.(5\%)$ What is the correlation coefficient between "motor$\_$UPDRS" and "RPDE"?

$5.(5\%)$ What is the average age of distinct subjects?

Mean age:

Median age:

Minimum age:

Maximum age:

Standard Deviation:

Py $($Numeric Data Type$)$

i$.$

iv$.$

Visualization $(25\%)$

$1.(8\%)$ Create a bar chart to display the distribution of subjects by gender.

o $(2\%)$ What is the gender ratio in the trial?

$2.(10\%)$ Create a scatter plot of "motor$\_$UPDRS" vs "age".

o $(3\%)$ Do you observe any trends or patterns between motor UPDRS

scores and age?

$(2\%)$ What potential insight could this give about disease progression?

Part $2-$ Feature Engineering $(30\%)$

Creating New Features:

$1.$ Jitter to Shimmer Ratio $(5\%)$

Create a feature called Jitter$\_$Shimmer$\_$Ratio by dividing Jitter$(\%)$ by

Shimmer. This will help analyze the balance between these vocal

features.

$2.$ UPDRS Progression $(5\%)$

o Create a feature called UPDRS$\_$Progression$\_$Rate, which captures the

rate of change in the "motor$\_$UPDRS" score relative to test$\_$time.

$3.$ Vocal Features Mean $(5\%)$

Compute the average of the Jitter$(\%),$ Shimmer, NHR$,$ and HNR values for

each subject. Store it as Vocal$\_$Features$\_$Mean.

Part $3-$ Predictive Modeling $(30\%)$

Linear Regression $(15\%)$

$1.(10\%)$ Fit a simple linear regression model to predict "motor$\_$UPDRS" using age,

gender, jitter, shimmer, NHR$,$ RPDE, and DFA as predictors.

$2.(5\%)$ Evaluate the model using Mean Absolute Error $($MAE$)$ and R$-$squared on the

test set.

Scatter Plot Analysis $(5\%)$

$1.$ Create scatter plots for "motor$\_$UPDRS" vs "Jitter$(\%)"$ and "motor$\_$UPDRS" vs

"Shimmer".

$(3\%)$ Do you notice any non$-$linear patterns in the relationship?

Non$-$Linear Relationships $(10\%)$

$1.$ Non$-$Linear Features $(15\%)$ : Generate polynomial terms $($e$.$g$.,$ square of the

jitter and shimmer$)$ to explore potential non$-$linear relationships. Use these

features in subsequent modeling.

$2.$ Once non$-$linearity is discovered, try fitting a polynomial regression model to

account for curved relationships between vocal features $($e$.$g$.,$ jitter, shimmer$)$

and motor UPDRS.

Column Description

name

subject# Unique identifier for cach subject which ranges from $1-31($Numeric Deta

Type$)$

age Age of Subject $($Numeric Data Type$)$

$5$ex Gender of Subject $($Categorical Data Type$)$

test$\_$time Time since recruitment into trial $($Numeric Data Type$)$

motor$\_$UPDRS Motor UPDRS score $($Numeric Data Type$)$

total$\_$UPDRS Total UPDRS score $($Numeric Data Type$)$

Jitter$(\%)$ Percentage of local variation in fundamental frequency $($Numeric Data

Type$)$

Jitter$($Abs$)$ Absolute jitter $($Numeric Data Type$)$

Jitter Difference between consecutive differences of fundamental frequency

$($Numeric Data Type$)$

Shimmer Difference between consecutive differences of amplitude $($Numeric Data

Type$)$

Shimmer$($dB$)$ Shimmer in decibels $($Numeric Data Type$)$

NHR Noise$-$to$-$harmonics ratio $($Numeric Data Type$)$

HNR Harmonics$-$to$-$noise ratio $($Numeric Data Type$)$

RPDE Recurrence period density entropy $($Numeric Data Type$)$