Applying Sentiment Detection Tools

The first step is to decide on what market sentiment the trader wants to tag or monitor. What emotion is being expressed relating to a market? This is where the rubber meets the road. The technology of text searches at this point in time is still a dumb technology. The search function is extremely fast in retrieving documents, but it is not that smart at filtering out the noise. A lot of unrelated documents get retrieved. This is because the Internet is full of unstructured text. It is a bag of words that has to be categorized.

So the binary option trader is really more advanced than the search engine, at least at specifying what to look for so the search engine retrieves the right documents. This is called semantic processing. Recently, IBM demonstrated a breakthrough in semantic processing and text search when Watson won a contest on the TV show Jeopardy. Watson, as you may recall, is the name of the IBM computer system that competed against two human competitors on Jeopardy. The breakthrough was that Watson wasn’t just searching its memory banks for keywords—it also had to understand how the words related to each other. In this era of Watson, computer programs will be important assistants in opinion mining and, as a result, it will mean any person is able to quickly understand the mood of the market. Until then, the trader has to do the work of a future Watson. But it can and should be done for binary option trading. Here is how. It is not that hard to do (Figure 3.3).