Recruitment Task – Classification of a Heart with Hypertrophic Cardiomyopathy (Cardiomegaly)

Task Description

Your task is to prepare a solution for the classification problem of detecting hypertrophic cardiomyopathy (cardiomegaly) based on the provided features.
You should use one of the classical machine learning methods described in the ML.md file.

The goal is to build a model capable of correctly distinguishing between a healthy heart and a diseased heart using the available data.

The repository contains a dataset stored in a CSV file, which includes selected geometric and imaging features used as the basis for further analysis.
We expect the candidate to prepare code that:

1. Loads the data from the task_data.csv file.
2. Splits the data into training and test sets.
3. Performs preprocessing (e.g., standardization, normalization, etc.).
4. Trains one or more selected models.
5. Evaluates the solution using cross-validation.
6. Evaluates the solution on the test dataset (e.g., using accuracy, precision, recall, F1-score).
7. Provides a brief description of the chosen approach.

---

Data (CSV)

The first column contains a photo ID, and the second column indicates whether the heart was diagnosed with cardiomegaly:

• 1 – positive diagnosis (diseased heart)
• 0 – negative diagnosis (healthy heart)

Below are the features describing the heart and lungs. Each feature includes its name (as used in the CSV file).

---

Lung width

The horizontal distance between the outermost points of the lungs.

Heart width

The maximum horizontal width of the heart.

CR ratio

The ratio of heart width to lung width.

Inertia tensors

Metrics describing the distribution of heart and lung pixels relative to the coordinate axes, capturing the shape and orientation of the objects.
The .csv file contains four components of this feature:

• xx – distribution of pixels relative to the y-axis (elongation along x)
• yy – distribution of pixels relative to the x-axis (elongation along y)
• xy – distribution relative to both x and y axes (a high value indicates object rotation)
• normalized_diff – a scalar value derived from the vector whose components are described above

Inscribed circle radius

The radius of the largest circle that can be inscribed within the heart area, describing its symmetry and compactness.

Polygon area ratio

The ratio of the area of the polygon enclosing the heart contour to the actual heart area.

Heart perimeter

The length of the heart contour.

Heart area

The area occupied by the heart.

Lung area

The area occupied by the lungs.

---

Expectations

• The code should be written in Python (e.g., using libraries such as scikit-learn, numpy, pandas, matplotlib).
• Include a short description of the solution in a Markdown (.md) file.
• Present evaluation results clearly (e.g., results table, ROC/PR curves).
• Code and commit messages should be written in English.
  The Markdown description may be written in either Polish or English.

To complete the task, please Fork this repository and submit the link to your fork in your recruitment form.

We recommend using Jupyter Notebooks for the implementation.

---

Example model

In the Example folder, you will find a fully implemented machine learning project that demonstrates the entire workflow — from data loading and preprocessing to model training, evaluation, and final conclusions. This example is designed to help beginner machine learning enthusiasts understand the structure of a complete project. It contains all the necessary information and practical guidance needed to successfully complete the recruitment task.

---

Evaluation Criteria

• Correctness of the code.
• Clarity and readability of the implementation.
• Use of classical machine learning methods.
• Creativity of the solution.
• Commit history in the repository – clarity and consistency will also be evaluated.

---

Good luck!