pip install pandas transformers sklearn nltk

Start the Jupyter notebook server

jupyter notebook

There are two files for initial mapping of Charts and Stocks datasets

1. Code/Para_Extracter.py - Extracting paragraphs from news dataset
2. Code/ParaPrice_Mapper.py - Map stock price chart with extracted paragraphs

Run these two files in order to create the mapped dataset, which will be used in all further notebooks

There are three notebooks:

1. Code/Twitter Stock volume extraction.ipynb - Extract recent tweets for a particular stock symbol.
2. Code/Twitter Breakout.ipynb - Detecting breakout hours from extracted tweets
3. Code/Web Content Extraction.ipynb - Extract news URL contents scraped from tweets

1. Code/Stock_Models.ipynb - Contains code for training the following algorithms: XGBoost, Decision tree, SVC, Naive Bayes, Random forest, Logistic regression
2. Code/gpt2classifier_recent_news.ipynb - GPT-2 classifier
3. Code/lstmcnnGlove.ipynb - LSTM-CNN with GloVe Embedding classifier

1. Code/Model earnings.ipynb - Logic for evaluating ROIs for any given prediction output

1. https://machinelearningmastery.com/sequence-classification-lstm-recurrent-neural-networks-python-keras/
2. https://stackabuse.com/python-for-nlp-word-embeddings-for-deep-learning-in-keras/
3. https://datascience.stackexchange.com/questions/45165/how-to-get-accuracy-f1-precision-and-recall-for-a-keras-model
4. https://towardsdatascience.com/text-classification-on-disaster-tweets-with-lstm-and-word-embedding-df35f039c1db