Large Scale Machine Learning Algorithm(s) for Stock Price Prediction
Summary
This project focuses on developing and testing complex stock analysis and prediction models utilizing state-of-the-art machine learning algorithms. While stock analysis and prediction may not be inherently intensive, this project seeks to expand upon presently accepted techniques in order to develop new and improved methods. The goal of this project is to integrate several databases (such as Google Trends, Tweets, general stock data, etc.) and conduct stock predictions with a large sum of data being fed into the models. Similarly, we seek to expand this project in a way such that we can conduct the analysis and predictions of our stock information in near-real time, offering a rapidly deployable prediction model. The results of this project will be shared in journals and conferences in order to add to modern techniques and approaches to time series analysis and predictive modeling.
Job Description
Student(s) will write programs to automatically download datasets, analyze datasets, and hopefully, write conference/journal paper(s) as the deliverable(s).
Computational Resources
This project will utilize multiple nodes concurrently to cut down computation time and leave ways for increased time for analysis and comparison. The programs will often be run several times over to discover discrepancies and/or inconsistencies. We may rely on XSEDE supercomputers to finish the tasks.
Contribution to Community
The results can be published in conferences/journals to bring a huge impact to the XSEDE community. The machine learning algorithms developed can be shared with the XSEDE community for scientific usage.
Position Type
Apprentice
Training Plan
We will train the student(s) for the basic machine learning algorithms, how to handle data analysis, and how to write a paper.