Over a million developers have joined DZone.
{{announcement.body}}
{{announcement.title}}

Using AI to Help Traders Behave More Rationally

DZone 's Guide to

Using AI to Help Traders Behave More Rationally

Recent research explores whether machine learning can be used to help detect undue emotion in traders and help them act (and trade) more rationally.

· AI Zone ·
Free Resource

Trading is an area of life in which automation has taken off like little other. While speed is undoubtedly a major selling point of robotrading systems, the ability to act rationally and devoid of emotions is also crucial to the success of such systems.

Despite robotrading taking off, however, human analysts are still prevalent in the financial sector, so it remains crucial to ensure that they act as efficiently as possible. Recent research from the University of Singapore explores whether machine learning can be used to help detect undue emotion in traders and help them act (and trade) more rationally.

The researchers use the unstructured text used to communicate the thoughts and opinions of investors to understand trading patterns and help traders make better investment decisions. They were able to use text data analytics to extract sentiment for certain topics from the analyst reports published on various listed companies.

Sentiment Analysis

The team used deep neural network supervised learning to analyze the reports at a sentence level rather than focusing on individual words. They believe that this gives them a better chance of extracting the true meaning of the text by understanding the context of each sentence.

They conducted this sentiment analysis on around 110,000 analyst reports written for companies listed on both the Tokyo and Osaka Stock Exchanges (the reports were written in Japanese). The findings from this were then incorporated into a second topic sentiment asset pricing model.

When the system was tested against other asset pricing models that typically did not incorporate sentiment analysis, they found that they were better able to predict the expected returns on a stock and were better able to understand which parts of the analyst report contributed to that stock movement.

"In our study, we found that topics reflecting the subjective opinions of equity analysts have greater predictability on portfolio returns than topics pertaining to objective facts and quantitative measures," the researchers conclude. "This seems to suggest that sentiment analysis could play a significant role in modern portfolio selection."

Topics:
trading ,sentiment ,emotion ,traders ,ai ,deep neural networks ,supervised learning

Published at DZone with permission of

Opinions expressed by DZone contributors are their own.

{{ parent.title || parent.header.title}}

{{ parent.tldr }}

{{ parent.urlSource.name }}