Applying Deep Learning to Financial Markets with News Data - PowerPoint PPT Presentation

Applying Deep Learning to Financial Markets with News Data

Executive Summary Large amounts of Use deep neural Predict market news data networks to build direction of predictive models relevant stocks

Dealing with News Original Workflow Find interesting Derive Make investment React in the articles implications decision market Limited scope of articles Very time consuming Missed opportunities

Dealing with News Future Workflow A+ Stream many news Link articles with Determine Determine if React in the articles appropriate stocks sentiment scores market moving market Able to determine sentiment Able to get in and out of Able to analyze more articles than and impact on the market market faster any human

Model Approach - Deep Learning Deep Learning cuDNN Library GPU abc NVIDIA Unstructured data NLP Lots of data Feature selection

Current Results Model Architecture Model Results Future Considerations Accuracy Tickers Financial Data Embedding Input LSTM $ 1 2 News Link to tickers GloVe Positive Neutral 3 4 5 A+ Sentiment Negative Softmax Output Batch 76% Normalization Market Movers F1 Score Proprietary Embedding

Thank You ? Questions