Do Tweets Help With Stock Prediction?
Directory of reports for stock prediction project.
Created on March 4|Last edited on March 17
Comment
What problem are we solving?
In this example project, we first explore ranges of hyperparameter values that produce an RL model with the highest sharpe ratio on legacy stock market data. We then explore whether information from Twitter can help the RL model learn to better predict those stock trends. This will improve the ability to make real-time, dynamic trading decisions, as well as allow for optimization of portfolio allocation.
Why is this important?
Machine Learning (ML) in finance forecasting has been recently gaining popularity as end-to-end models are starting to outperform traditional methods. However, the problem of continual learning, where the model updates itself constantly as new data comes in, is still a major problem for ML approaches. Reinforcement Learning (RL) is an ML approach that has been the most successful at solving the continual learning problem in other contexts, but until recently was not applied to finance prediction problems.
Weights and Biases makes it easy to compare models, so we can immediately assess the outcomes of changing the training data.
This landing page breaks down the findings of this project for multiple different audiences role types within a FinTech organization that would be interested in the methodology, progress, and results.
For the Practitioner
Due to the complex nature of RL algorithms, it is important to monitor a variety of metrics and state variables to understand how the agent (in our case, trader) is interacting with the world (in our case, the market). W&B makes this easy by providing a clean view of metrics that update live during training. The reports below are crafted for a technical audience who may want to emulate the findings from this project. They contain a granular look at the methodology and resultant findings.
For the Team Lead
This report is built to give a team lead or technical manager a view into how their teams are progressing on the project. While still geared towards a technical audience, this report is more succinct and highlights key findings that will allow a team lead to quickly understand if their team is making progress. Key metrics and KPIs that will allow a lead to determine if the methodology behind the new findings is feasible for production are captured here.
For the Stakeholder
A non-technical audience or business stakeholder will certainly be interested in how the technical work their organization is doing will ultimately improve trading strategies and impact the bottom line. The following report shapes the findings of this project in a way that does not require technical acumen - it simply communicates how these new proposed trading strategies are going to provide a positive impact on the business.
Add a comment