History Article
Close
This entry was posted on January 4, 2016 at 6:37.
The last change is the January 4, 2016 at 14:00.
Friday, 2 October at 14.30. The US Department of Labor publishes employment data. The number is lower than expected: in a flash the yield of ten-year US government bond falls by more than 2% to 195 basis points. Wednesday, December 3 last year. Always at 14.30. Mario Draghi, ECB president, reveals the moves on the strengthening of quantitative easing. The strategy disappoints and, in a flash, the euro against the dollar spikes,: 1,055 to get to share 1.07. Of similar examples, old and new, the markets are full. And, in future, continue to be more and more. Is it any wonder? Of course not. It is the effect of the algorithms, especially those ultra-fast, they do their job: they react instantly to the news; exploit the many folds of the lists.
You say there is no real news. The High frequency traders (HFT), however, a bit ‘subdued after recent tight regulations, are a known phenomenon. True! Yet there is one aspect which is discussed shortly. The Hft is in fact only one aspect of a broader dynamic: the automated trading. Which, on the one hand, it is not characterized by the speed of execution. And, second, it is a more complex and deep trend that makes its way into finance. An approach in which man tends to end up in the background. At center stage, more and more, there are robots.
The proof? The provides AiteGroup. According to the research firm, algorithmic trading on the stock market, in 2015, managed the 66% of trade. In currencies, then, the volumes in the hands of the software have been 27%. And what of the traditional bonds: here the software have exceeded the threshold of 10%. In short, the importance of the phenomenon is in the numbers.
At that begs the question: what are the technologies used by such investors? “The strategies-meets Petter Kolm, director of the department of financial mathematics at the University of New York – are different. Hedge funds use techniques such as, for example, statistics, econometrics or machine learning to search, or “build” signals. ” Indications to be exploited, then, in algorithmic trading.
Yes, the operational guidelines. These in hindsight are pursued more and more not only in the markets and the prices of securities. But, in the vast sea of digital information on the network: from those of the agencies to the tweets on Twitter. “So-indicates Fabrizio Lillo, professor of financial mathematics at the Scuola Normale in Pisa -i software becomes very useful. The system, in fact, are able to understand the texts and, therefore, provide any signal to traders automatically. ” As it happens? “First, the algorithm needs to understand what it’s about. So, for example, the ticker of individual actions allow to identify the company involved in the news. ” After that, the software sets up the polarity of the information. That is, it determines whether the news is positive, neutral or negative. “At this level comes in the so-called classifier. The system, based on its dictionary that contains the domain of many words, is able to identify the words. Then, by calculating the amount of positive and negative terms, identifies the polarity of the news itself. ”
But it is not just a matter of statistics. “There is also the semantic algorithm – says Marco Varone, President of Expert System -.I our software, in fact, are able to identify the relationship between two different words. It is a “capacity” resulting in the creation, over the years, a large knowledge base stored in which the words have been correlated between them. ” So the algorithm, first, performs parsing (distinguishes and identifies, for example, the names and verbs). Then, he realizes that logic. Finally, “it undertakes the analysis on the relationship of words.” The interpretation provided the message is sent to the robot. Which, in turn, take up the dances. With all due respect human investor.
© ALL RIGHTS RESERVED
Permalink
No comments:
Post a Comment