Analysis and Classification of Parkinson's Risks and Symptoms

Parkinson Risk and Symptom Analysis and Classification

Context

Parkinson's disease is a progressive neurodegenerative condition that affects movement control and cognitive functions. It frequently occurs in people over 60 years old and manifests through a combination of motor symptoms (tremor, rigidity, bradykinesia, postural instability) and non-motor symptoms (sleep disorders, constipation, depression, speech problems).

Early diagnosis and monitoring of disease progression are essential for optimizing treatment and improving patients' quality of life. Artificial Intelligence enables complex analysis of clinical data and lifestyle for:

• identifying risk factors
• classifying patients
• estimating disease progression

To solve this problem you will have two files at your disposal:

• train.csv – contains all variables, including Diagnosis, for training models
• test.csv – contains the same variables, without the Diagnosis column, and the data from this file is used for generating predictions

Dataset Description

Requirements

For this problem you have to solve 3 tasks.

Task 1 – Cardiometabolic Risk Score (20 points)

Calculate for each patient in the test set a cardiovascular and metabolic risk score based on existing columns (Hypertension, Diabetes, BMI).

• Formula:

1CardiometabolicRiskScore = (Hypertension == 1) + (Diabetes == 1) + (BMI > 30)

CardiometabolicRiskScore is a simple score that evaluates the patient's cardiovascular and metabolic risk.
Each component adds one point to the score:

• Hypertension: if the patient has high blood pressure, 1 point is added.

• Diabetes: if the patient has diabetes, 1 point is added.

• BMI > 30: if the body mass index indicates obesity (BMI over 30), 1 point is added.

• Answer in the submission file will be the calculated numeric value for each patient in the test set.

Task 2 – Lifestyle Risk Index (20 points)

Calculate a score that evaluates lifestyle-related risk based on existing columns (Smoking, AlcoholConsumption, PhysicalActivity).

• Formula:

1LifestyleRiskIndex = (Smoking == 1) + (AlcoholConsumption > 2) + (PhysicalActivity < 1)

LifestyleRiskIndex is a score that evaluates the risk associated with the patient's lifestyle.
Each component contributes one point to the score:

• Smoking: if the patient smokes, 1 point is added.
• AlcoholConsumption > 2: if the patient consumes more than 2 units of alcohol per week, 1 point is added.
• PhysicalActivity < 1: if the patient performs less than one hour of physical activity per week, 1 point is added.

The final score varies between 0 (risk-free lifestyle) and 3 (presence of all risk factors), providing a simple measure of lifestyle impact on health.

Task 3 – Parkinson Classification (60 points)

Build a classification model that predicts Diagnosis (0 = healthy, 1 = Parkinson's) based on all available information in train.csv. You can choose to use all information from the training set for model classification or you can select the most appropriate information.

• Predictions for each patient in test.csv must be in 0/1 format.
• Answer in the submission file will be 0 (healthy) or 1 (suffers from Parkinson's) for each patient.

Submission File Format

The submission.csv file must contain:

• PatientID - represents the patient ID and is taken from the test.csv file;
• subtaskID can be: Task1, Task2, Task3
• Answer represents the calculated value for the patient indicated by PatientID for the task indicated by subtaskID.

Evaluation

For subtask 3, the score is awarded according to the following rules:

• AUC ≥ 0.95 → 60 points
• AUC ≤ 0.70 → 0 points
• The interval between them: proportional score

For subtasks 1–2, exact evaluation is performed (by comparison).