Before you start using scripts in this repository you have to configure your workspace. We've limited ourselves to only
use Python 3. So you have to be sure that you've installed it. Some of the requirements are listed in requirements.txt
file. These you can install with pip
or conda
. Moreover:
- make sure that all the data downloaded from Kaggle are unpacked in
./data/raw
directory, - the information about GloVe model that we've used in modelling can be found here,
- for running RNN models you have to install cuDNN or replace
CuDNNGRU
layers with plain oldGRU
(slower).
All scripts should be run from the repository's root directory!
- Download and unzip all raw data from Kaggle to
./data/raw
. - Run
./scripts/data/change_classes.py
to assign labels to dataset based on stock open and close value. - Prepare data used for corpus cleaning by running
./scripts/data/prepare_*.py
and than clean corpus with./scripts/clean_corpus.py
. - If you intend to run a GloVe or Doc2Vec model you must first run a proper script from
./scripts/features
directory. - Now you should be able to run scripts from
./experiments
and./scripts/model
directories.