Predict which startup will earn the most profit using deep learning and machine learning techniques!
This project empowers entrepreneurs and analysts to make data-driven decisions by modeling real-world financial data from 50 different startups across the USA.
- Goal: Build a high-performing neural network to predict profit for a startup based on its investments and location.
- Dataset: 50_Startups.csv — includes R&D Spend, Administration Spend, Marketing Spend, State, and Profit.
- Tech Stack: Python, Pandas, NumPy, Scikit-learn, TensorFlow (Keras)
-
Deep Learning Model:
Utilizes a multi-layered neural network for powerful regression predictions. -
Automated Early Stopping:
Custom callback monitors validation score for optimal model training. -
Advanced Data Preprocessing:
One-hot encoding for categorical variables, robust and min-max scaling for features and labels. -
Interactive Prediction:
Use your own startup data to get instant profit predictions!
-
Data Preparation
- Cleans and preprocesses the startup dataset for optimal learning.
- One-hot encoding for the state/location feature.
-
Model Training
- Scales features, trains a deep neural network.
- Early stopping to prevent overfitting.
-
Profit Prediction
- Enter R&D, Administration, Marketing spends, and State interactively.
- Instantly retrieve the expected profit for your startup.
Requirements
- Python 3.7+
- Pandas, NumPy, scikit-learn, TensorFlow
To Run:
- Clone this repo:
git clone https://github.com/amdollar/50_Startups.git cd 50_Startups - Install dependencies:
pip install -r requirements.txt
- Open and run
50_Startups.ipynbin Jupyter Notebook or VSCode.
- Predict profit for a new startup in New York with your own investment figures!
- Explore feature scaling and neural network design for regression tasks.
- Entrepreneurs: Optimize your investment strategy.
- Students: Learn deep learning best practices for tabular data.
- Data Scientists: Extend and experiment with scalable regression pipelines.
Contributions welcome! Open issues for questions, ideas, or improvement suggestions.
Unleash the potential of your startup — predict, invest, grow! 🚀