QRNGClassifier

The following is a collection of Python scripts and data files used for classifying and predicting what quantum device/computer input strings of quantum generated random binary raw data (QRNG) came from. Simply predicting by chance, you would get roughly. around 25% accuracy if the data is generated by 4 different QC's (which our dataset), but our model (specifically using a gradient booster classifier model with a few more data processing features such as entropy calculations and Hadamard linear spectral transform) was able to reach >78% prediciton accuracy for classifying QNRG strings to the quantum machines that generated them by mapping quantum noise and bias.

The following is a list of the steps we used to constuct out final model:

1. Obtained raw binary output QNRG from DoraHacks repo with 3 different QC outputs (6000 lines, 100 bits each)
2. Formatted and inputted AWS Rigetti QC output data (raw form output data from DoraHacks repo)
3. Combined datasets and processed data by creating/formatting a dataframe, breaking up binary QNRG strings, concatenating lines by groups of 4, applying min/shannon entropy measures, and finally applying hadamard linear spectral QFT transform.
4. Broke data up into training/testing datasets using sci-kit learn and ran formatted/feature-processed df through gradient booster model