Big Data is much more than the sum of its parts. It is a whole world of possibilities that represents a constant challenge to decipher its mysteries, to take advantage of its possibilities in the most multiple fields, including those related to stock market predictions.
Specifically, the synergistic power of Wikipedia and Google, two giants of the web, could have great potential to make revolutionary economic predictions to that end. So affirms it, at least, the still nascent science of big data sets, that in his short life he has already reaped enormous successes, but it has also contributed resounding failures, as just happened with Google's misleading Flu Trends reports on trends in influenza and similar illnesses.
The investigation comes a few months after the scam of this report, with the addition that economic predictions are somewhat complicated, especially in such a changing and elusive terrain as the stock market. Even so, the conclusion of the work is clear: it is feasible to predict the moment when financial markets will tend to have a fall. Of course, there is still a long way to go to achieve it.
Target: detect valuable information
A group of researchers from the University of Warwick, in the United Kingdom, claims to have demonstrated that it is feasible to intelligently use the data provided by Wikipedia and Google to carry out a joint analysis that leads to the long-awaited results.
Discover trends, the Holy Grail of Big Data analytics, in this circumstance it is translated into the use of both data sets applying a modus operandi which is based on the appearance of phrases in Wikipedia associated with previous crises Actions, and combining these results with the search terms de Google.
If progress is made on this path, the logic of mistrust over the null reliability of economic forecasts would end. “I don't read economic forecasts. I don't read comics”, says Warren Buffet jokingly for decades.
“We use a known technique of Computational linguistics to identify semantic themes within this corpus, what is Wikipedia ", the researchers explain. From this point, they had the help of Mechanical Turk to apply human intelligence and name each of these lists obtained.
In conclusion, examined the incidence of these problems in Google searches over the period 2004-2012, based on Google Trend search engine, not in actual queries, trying to detect relationships between these searches and market movements.
It's still a distant dream
The investigation revealed that there is a correlation among the things that people look for in the Google search engine before making a purchase-sale choice. This means that this type of behavior could also be found in other types of scenarios..
In other words, research suggests that “extensions to these analyzes could provide insights into the large-scale flow of information prior to a series of real-world events.
“In complex events, like movements in financial markets, valuable information can be contained in search engine data using keywords, but with less obvious semantic connections”, study concludes.
In other words, still is An utopia Investors start using Google to shape their stock market strategies. As long as Big Data does not reliably prove otherwise, it will be necessary to continue agreeing with Mr. Warren.
Image source: cooldesign / FreeDigitalPhotos.net
Related Post:
