Trade Exception-triggered Short-term Stock Trend Prediction based on Event Detection and Sentiment Analysis

The project aims to establish a short-term trend forecasting mechanism for Hong Kongand Shenzhen stock based on event detection and sentiment analysis techniques. Wecollect data from multiple sources inside and outside Hong Kong and use deep neuralnetwork and data reconstruction technology for effective event detection and modeling,as well as analyzing the public sentiment towards the events, so as to provide the publicinvestors with more comprehensive information and knowledge on related events. Thisnot only can facilitate them to form relatively rational and objective judgment on recentmovements of the underlying stocks, but also may help investors from Shenzhen andHong Kong to reduce investment risks caused by information asymmetry. It is expectedthat our project can contribute to promote cooperation and development of financialmarkets and technologies between Shenzhen and Hong Kong.