2024 · ML Engineer / Developer

Cocoa Stock Price Prediction

Discipline

Machine learningCoding

For

Study

Stack

PythonFlaskTensorFlowStreamlitLSTMFeature EngineeringTime Series Analysis

Links

GitHub

Achieved 90% forecasting accuracy with MSE of 0.00048 and MAE of 0.0215 on cocoa market data

What went into it

• Engineered a cocoa market price prediction model using an LSTM (Long Short-Term Memory) neural network for time series forecasting of commodity prices.
• Employed advanced feature engineering techniques to extract meaningful temporal patterns from historical cocoa price data, significantly improving model performance.
• Achieved 90% forecasting accuracy with an exceptionally low MSE of 0.00048 and MAE of 0.0215, demonstrating production-viable prediction capabilities.
• Deployed the model via a Streamlit web interface backed by a Flask API, enabling interactive exploration of predictions and historical trend analysis.