Machine Learning with SGD Regressor and Random Forest Classifier
This project applies two machine learning models: SGD Regressor and Random Forest Classifier. The goal is to understand how each model performs in practice and when it's a good idea to use them.
About the Models
Features that are needed for the model to predict : no, Brand, CPU, GPU, RAM Type, RAM Size, Storage, Screen Type, Refresh Rate, Uniqueness, Upgradeable, price.
You can see an example of the dataset that is accepted by the model here
DatasetSGD Regressor
Stochastic Gradient Descent (SGD) is a fast and efficient way to train linear models. It is great when:
- Your data is large or comes in streams.
- You want something fast and memory-efficient.
- You need to add regularization (like L1 or L2) to prevent overfitting.
You can try this model rightaway by downloading this
SGD RegressorRandom Forest Classifier
Random Forest is an ensemble learning method that builds many decision trees and combines their results. It works well for most classification problems out of the box and is known for being:
- Accurate and robust, even when data is messy
- Resistant to overfitting
- Easy to understand and implement
You can try this model rightaway by downloading this
Random Forest Classifier
Why Use These Models?
We chose these models because they balance simplicity and performance. SGD Regressor gives us fast, scalable predictions for regression, while Random Forest handles classification with high accuracy and reliability. Together, they’re a strong starting point for exploring a variety of machine learning tasks.
Setup
1. Installation
Make sure you're using Python 3.8+ and working in a virtual environment (recommended).
1# 1. Create a virtual environment (recommended)
2python -m venv venv
3source venv/bin/activate # Linux/macOS
4venv\Scripts\activate # Windows
5
6# 2. Install required libraries
7pip install -r requirements.txt
8
2. Usage
Here’s a basic example of how to load and use the models:
1import pandas as pd
2from joblib import load
3
4# Load trained models
5sgd_model = load("models/sgd_regressor.pkl")
6rfc_model = load("models/random_forest_classifier.pkl")
7
8# Load test data
9X_test = pd.read_csv("data/X_test.csv")
10
11# Make predictions
12predicted_price = sgd_model.predict(X_test)
13predicted_category = rfc_model.predict(X_test)
14
15print("Predicted Price:", predicted_price[0])
16print("Predicted Category:", predicted_category[0])
17
Ensure the model files (.pkl) and input data (X_test.csv) are in the correct directories.
Our Team
Member :
- Glenn Josia Devano (glenn.josia.devano@ui.ac.id)
- Nadhim Muzhaffar Hakim (nadhim.muzhaffar@ui.ac.id)
- Zayyan Ramadzaki Firdaus (zayyan.ramadzaki@ui.ac.id)
Coba Coba Berhadiah
Building this model is a submission for Us to join Datathon 2025 Competition.
