Fintech Application

Market News Analysis and Accelerated Stock Price Prediction

Explore More

About the Project

Whether there is a pattern behind the ups and downs of the stock market has always been a controversial topic. On one hand, the technician looks at historical price charts to make decisions. On the other hand, fundamental analysts utilize data from various sources. In particular, financial news is a vital asset considering the vast amount of it circulating in our daily life. Our goal here is to utilize Natural Language Processing on the textual data of news, which further prediction is based on using multiple algorithms from Machine Learning, and also to improve the existing computational efficiency.

What we do

Data Retrieval

Collect data, including market news and stock price, from online sources automatically.

Data Preprocessing

Implement Natural Language Porcessing Techinques to extract useful data from the data set.

Feature Identification

Analyze the preprocessed dataset, and identify the features which affect the stock price more significantly.

Stock Price Prediction

Predict the stock price based on the models that are constructed after the learning process.

Computation Distributing

Implement Parallel computation Techinques to distribute the price prediction workload

Prediction acceleration

Accelerate the stock prediction process via parallel computation on GPU.


Architecture Design


Our team

Dr. SM Yiu


Ru Chiang

Team member

Justin Wang

Team member

Ray Huang

Team member

Drop us a line


The University of Hong Kong, Pokfulam, Hong Kong