This repository contains an end-to-end workflow for predicting option price returns over a five-day trading period. The model leverages a rich dataset that combines historical options data, fundamental metrics, institutional activity, liquidity indicators, and sentiment signals.
-
Goal: Predict 5-day returns on options prices using a combination of fundamental, institutional, liquidity, and sentiment data.
-
Data Sources:
- Historical options data (via ThetaData)
- Various fundamental and technical data
- Institutional ownership and sentiment indicators
- Clone the Repository
git clone https://github.com/themoonwalker1/quantcap-options.git
- Generating a New Dataset:
cd quantcap-options/dataretrieval- Open and run options_data_retrieval.ipynb
- Output will be a csv file, similar to final_dataset_5_trade…
- Accessing historical options data, requires ThetaData subscription and terminal. More information found here: https://http-docs.thetadata.us/
-
Navigate to Model Directory
cd to quantcap-options/final_model
-
Run Model Notebook
- options_ml_predicting_price.ipynb is the primary notebook
- Default dataset used: final_dataset_5_trade_days.csv
- Adjust hyperparameters and preprocessing steps as needed
- Notebook outputs performance metrics and visualizations
-
Attempted Models
cd quantcap-options/attemptedmodels- To explore experimental models and alternative approaches