Project information

• Category: Data Science & Machine Learning
• Programming Ecosystem: Python (TensorFlow/Keras, scikit-learn/sklearn, Scipy, NumPy, Pandas, Matplotlib, Plotly)
• Project date: Winter 2022 - Present
• Project URL: Sample

• Implemented various statistical methods and advanced machine learning models including Bollinger Bands, Support Vector Machines (SVMs), and Long Short-Term Memory (LSTM) networks to analyze and predict stock market trends
• Utilized a comprehensive tech stack including Keras, Sklearn, Scipy, and NumPy for model development, alongside Matplotlib and Plotly for effective data visualization
• Conducted in-depth research into parameter optimization, reading several research papers and exploring different combinations of technical indicators
• Experimented with prediction outputs, examining price, change percentage, and direction as potential options and fine-tuning minimum confidence thresholds
• Investigated stock clustering techniques based on research paper to group similar companies for more robust model training, learning about winsorizing of outliers in the process