### Task Description: Predicting Click-Through Rate (CTR) for Targeted Advertising

#### Problem Statement
The objective of this task is to predict the **weighted Click-Through Rate (wCTR%)** for targeted advertisements. Each row in the dataset represents an advertising instance with features describing the ad group, format, platform, device type, targeting category, demographics, and other contextual details. The goal is to leverage these features to build a model that accurately predicts the CTR performance for ads.

#### Dataset Details
- **Training Dataset:** Contains 13 columns and 9,661 rows. Each row represents an advertising scenario with features that include:
  - **Группа:** Ad group, which may indicate the campaign target.
  - **Формат:** Ad format (text or graphical).
  - **Тип площадки:** Platform type (e.g., search, social).
  - **Тип устройства:** Device type (e.g., mobile, desktop).
  - **Пол:** Targeted gender.
  - **Возраст:** Targeted age group.
  - **Категория таргетинга:** Targeting category (e.g., competitor-based, interest-based).
  - **Уровень платежеспособности:** Consumer purchasing power level.
  - **Упоминание брендов:** Whether the ad mentions a brand.
  - **wCTR (%) > 0:** Target column indicating the weighted CTR percentage > 0.

- **Test Dataset:** Contains similar columns, except the target column (`wCTR (%)`) is missing, and predictions are required.

#### Submission Format
The solution file must:
1. Contain an **"id"** column to uniquely identify each row.
2. Include a **"prediction"** column with the predicted wCTR values.

#### Objective
Develop a machine learning model that:
- Handles categorical and numerical data effectively.
- Predicts the CTR (`wCTR (%)`) based on the given features.
- Provides a reliable estimate of advertising success across different audience segments and ad types.

This challenge tests the ability to work with real-world advertising datasets, preprocess data with categorical variables, and build robust predictive models.