Andreas Polycarpou – How Does Algorithmic Trading Works? | Methodology & Sentiment Analysis

Andreas Polycarpou further explains how Algo Trading comes to play essential roles in trading the financial markets. Here are some basic facts on the methodology of Algo Trading and how the implementation of Sentiment Analysis affects trading.

Methodology- what must be done?

Andreas Polycarpou says, “The recent backlash of Facebook is a development that will affect stocks. However, with the proper understanding of Sentiment Analysis, stock trends can be mapped easily. The first step is to crawl articles on leading networks such as Business Times for information related to Facebook. After which you mine the texts and derive relevant sentiment per day and buy or sell 10 shares of its stock, either the sentiment increases or decreases by 0.5. It is noteworthy that today’s trading is based on yesterday’s sentiment. And sentiment ranges from -1 to 1 with 0 signifying neutral.”

Sentiment Analysis- what is the VADER approach?

Andreas Polycarpou explains, “VADER stands for Valence Aware Dictionary for sEntiment Reasoning, which is a pre-built model packaged in Natural Language Toolkit (NLTK). It can give both positive/negative (polarity) as well as the strength of the emotion (intensity) of a text. It is rule-based. It relies heavily on humans rating texts via Amazon Mechanical Turk — a crowd-sourcing e-platform that utilizes human intelligence to perform tasks that computers are currently unable to do. This means that other people have already done the dirty work of building a sentiment lexicon for us. These are words or any textual form of communication generally labeled according to their semantic orientation (as either positive or negative) for us.”

“The sentiment score is obtained by adding up the intensity of every word that makes up the text and then normalizing it. Here are the 5 heuristics use for rating VADER with typical examples:

  • Punctuation– I love money versus I love money!!
  • Capitalization– I’m thirsty versus I’M THIRSTY
  • Degree modifiers (engagement of intensifiers) – I WANT MONEY!! VS I REALLY WANT MONEY!!
  • Conjunctions (shift in sentiment polarity, with further dictating polarity)- I love money, but I really hate illegal money
  • Preceding Tri-gram (denoting reverse polarity through tri-gram in a lexical feature) – Illegal money is not really all that great.

With Financial News, Sentiment Analysis, and other relevant indicators, building a perfect Algo Trading strategy becomes easy for high-end performance.” Concludes Andreas Polycarpou.

Be the first to comment

Leave a Reply

Your email address will not be published.