Intelligent sentiment analysis systems for detecting and estimating the impact of sentiment on market movements

Prices of company shares, national currencies, and commodities change suddenly usually on the whim of people who buy and sell shares or currencies or commodities. This leads traders to buy when they should sell, and to sell when they should buy – the whim or the sentiment of single trader can create a boom or bust in financial and commodity markets. There is even a subject dedicated to these whims called behavioural finance that studies the impact of sentiment on market prices. Sentiment analysis is based on algorithms that help analysts to relate prices to sentiments. These algorithms are based on work in Artificial Intelligence (and in Econometrics), which help to estimate (loosely predict) the impact of sentiment. AI-based algorithms can be implemented by natural language processing technologies for analysing texts and through event-driven technologies.

You will be looking at established markets in the USA and Europe together with newer markets in especially in China.

In this project you will examine and implement the algorithms used to extract sentiment and estimate the impact of sentiment on financial markets. Such systems are in great demand in finance, agriculture and food industries, and many others.

See (i) a recent paper on high-frequency trading and the mayhem caused by Gamestop share prices (https://www.dropbox.com/s/osgvc3xur3zyodz/2021_Future%20Tech%20Conf_Market_movements_at_high_frequencies_and_latency_in_response_times__FTC.pdf?dl=0)

(ii) my piece on Brexit and Ireland ( https://www.rte.ie/brainstorm/2018/0427/958625-how-sentimental-are-we-about-brexit/ ); and

(iii) a paper on the impact of sentiment on oil markets (attached)(https://dl.acm.org/citation.cfm?id=3222293)