Exam Score Prediction - OJIA 3 Simulation

🏫 Exam Score Prediction

For this problem you need to implement a regression model capable of predicting the exam score (Exam_Score) using an available dataset. The dataset is organized in a CSV file, and the model's performance will be evaluated based on Mean Absolute Error (MAE).

---

🔹 Dataset

The dataset contains the following columns:

• ID: Unique student identification number
• Hours_Studied: Number of hours spent studying
• Attendance: Percentage of class attendance
• Parental_Involvement: Level of parental involvement (Low, Medium, High)
• Access_to_Resources: Access to educational resources (Low, Medium, High)
• Extracurricular_Activities: Participation in extracurricular activities (Yes/No)
• Sleep_Hours: Average number of hours of sleep per night
• Previous_Scores: Previously obtained exam scores
• Motivation_Level: Motivation level (Low, Medium, High)
• Internet_Access: Internet access for studying (Yes/No)
• Tutoring_Sessions: Number of tutoring sessions
• Family_Income: Family income (Low, Medium, High)
• Teacher_Quality: Teacher quality as perceived by the student (Low, Medium, High)
• School_Type: Type of school (Public/Private)
• Peer_Influence: Peer influence on the student (Negative, Neutral, Positive)
• Physical_Activity: Number of hours of physical activity per week
• Learning_Disabilities: Presence of learning disabilities (Yes/No)
• Parental_Education_Level: Parents' education level (High School, College, Postgraduate)
• Distance_from_Home: Distance between home and school (Near, Moderate, Far)
• Gender: Student's gender (Male/Female)
• Exam_Score: Target variable, exam score obtained (numerical value)

---

📚 Tasks

Subtask 1 (10 points)

Calculate the mean of values from the Hours_Studied column on the training set. For each student in the test set, determine the absolute difference between Hours_Studied and the calculated mean.

Subtask 2 (10 points)

Determine for each student in the test set whether they sleep little (<7 hours). The result will be True or False.

Subtask 3 (10 points)

Count how many students from the training set had a previous score (Previous_Scores) greater than or equal to each student in the test set.

Subtask 4 (10 points)

Determine the number of students from training with the same motivation level (Motivation_Level) as each student in the test set.

Subtask 5 (60 points)

Build a regression model to predict Exam_Score based on the provided features. The model must generalize to new data and will be evaluated with MAE.

---

📊 Notes about the dataset

• Target field: Exam_Score
• The model must be trained on train_data.csv and evaluated on test_data.csv
• It is recommended to check for missing values and normalize features.

---

⚙️ Evaluation criteria

• Performance: MAE as low as possible

---

📨 Submission file format

The submission file must be a CSV with exactly three columns:

Column	Type	Description
subtaskID	integer	Represents the subtask ID (from 1 to 5).
datapointID	integer/string	Represents the unique identifier of the row from the test set (ID).
answer	float / int / bool	Answer for the respective subtask. The value type depends on the subtask: • Subtask 1: float • Subtask 2: boolean • Subtask 3: integer • Subtask 4: integer • Subtask 5: float (model predictions)

🔹 Examples

subtaskID	datapointID	answer
1	101	12.5
2	101	True
3	101	7
4	101	3
5	101	85.3

Important: Each row in the CSV represents the answer for a single subtask and a single datapoint. For each datapointID there must be one row for each subtask.

---

📝 Note

Submitting a sample_output generates 6 points.

🗂️ Useful resources

• Complete Starter Kit – contains a skeleton from which you can start solving the problem