In this repository we published our engineering thesis (PDF) alongside code used. Thesis_codes directory contains all of our calculations and plots of our engineering thesis "Discovering Impact of Market Microstructure on Stock Prices by Computational Intelligence Approach".

First one called "Discovering_market_microstructure_parameters.ipynb" presents implementation of market microstructure parameters and show dependencies between them.

Second one called "Experiments with ML models and LSTM.ipynb" presents implementation of Machine Learning models and LSTM as well as experiences described in Chapter 5 and 6 in order to find optimal set of parameters in predicting maximal swing on price.

Last one called "skipping_params_MLP_and_th_selection.ipynb" presents process of choosing optimal market state hyperparameters, finding optimal set of market mirostructure parameters and experiences described in Chapter 5 and 6 of MLP model.

We used few python libraries such as: NumPy (Version 1.20.2), Pandas (Version 1.2.3), Matplotlib (Version 3.2.2), Pytorch (Version 1.10.1) and Tensorflow (Version 2.7.0).