soni0391s/game-sales-prediction-model

Game Sales Prediction Model

This model predicts Global Sales of video games based on several features.

🔢 Input Features

Numerical:

• NA_Sales
• EU_Sales
• JP_Sales
• Other_Sales
• Year

Categorical:

• Platform
• Genre
• Publisher

🧠 Model

• Trained using scikit-learn
• Algorithm: KNeighborsRegressor
• Preprocessing: StandardScaler + OneHotEncoder (ColumnTransformer)

📦 Files Included

• game_model.pkl → Trained ML model
• inference.py → Script to load and run predictions
• requirements.txt → Dependencies

🚀 Inference (example)

from inference import predict

sample = {
    "NA_Sales": 2.5,
    "EU_Sales": 1.8,
    "JP_Sales": 0.5,
    "Other_Sales": 0.3,
    "Year": 2010,
    "Platform": "PS3",
    "Genre": "Action",
    "Publisher": "Sony"
}

print(predict(sample))