Using Sentiment Analysis for a new level of insight.
Sentiment allows a user to measure the impact of a particular news item or news flow around an instrument and evaluate its potential impact, in real-time. Compiling data from traditional news sources and social media, has led to the necessity for new real- time analytical tools, to help users filter and understand the explosion in data.
Sentiment analysis, however, is not merely categorizing sentiment in positive, negative or neutral terms. The intensity of an event’s impact is also influenced by the authenticity of the source and the weighting associated with it. Marketware’s Strategic partner Heckyl has done 5 years of back testing to rate all sources with regards to their accuracy across news sources, bloggers and tweeters, on a sector and individual company basis. This back testing allows us to filter and curate the sources to ensure that only relevant and impactful news is consumed and analyzed.
We believe that it is important to weight sentiment towards traditional news sources, rather than social media such as twitter and we currently work to a 90/10 split. From the analysis that we have done once you are researching instruments that are not large caps there is not enough content on twitter to generate a meaningful sentiment value. Therefore the abundance of content from main line sources such as national, local news and trade publications online provide the breadth of coverage needed, especially once you start looking at the smaller sized companies.
Sentiment data is used by a number of our clients in differing ways.
Direct correlation to the underlying share price varies across market cap and sector, but we have found a high correlation to large cap stocks. In addition a number of clients use sentiment as a leading indicator and others are starting to use it as another technical indicator that needs to be considered. Finally a number of clients see sentiment as a key filtering mechanism to focus on the most positive and negative instruments, where more research or specific action may be required. Ultimately it provides a quick way to digest all of the news flow around a company and take the appropriate action. In a world of exponentially increasing data and news this will only become more valuable for users with less time.
Unleashing the power of Artificial Intelligence (AI).
Underlying this all is a machine learning, natural language programming and AI system that are constantly improving. It is continually analyzing the web to ensure that we are covering the most relevant sources and dropping those that no longer reach the required mark, as well as finding and adding new sources as they come online. It is also constantly learning from the new inputs and market reaction to improve the algorithms and sentiment score.