Software developed at Oxford University that accurately assesses what people mean from what they say online will provide a valuable ‘sentiment analysis’ tool for businesses, particularly finance companies.
A new company, TheySay, has been set up by the University’s technology transfer company, Isis Innovation, to exploit the software that is based on work by Professor Stephen Pulman and Karo Moilanen from Oxford University’s Department of Computer Science. It has received first-round investment of up to ￡500,000 from IP Group plc.
Understanding sentiment is a proven source of competitive advantage for businesses, public sector bodies, political organisations, and individuals keen to monitor and measure what is being said about them on the Internet.
TheySay’s approach of using linguistic intelligence along with machine learning techniques overcomes the limitations of existing sentiment analysis tools. The researchers used an exhaustive sentiment classification and scoring scheme to classify linguistic structures for sentiment and to score all individual entity mentions in text.
Professor Stephen Pulman, of Oxford University’s Department of Computer Science, said: ‘We have a very large database of words annotated by hand along several dimensions for the emotional meaning they carry, and we also evaluate the grammatical context in which these words occur, taking account of the effects of negation and other constructs that change meaning. A word like 'progress' is generally perceived as positive, but not when it is in a context like 'fail to progress', or 'little progress'.
‘By taking account of grammatical context we can determine emotional attitudes towards the entities and relations mentioned in a text.
‘Many companies are concerned about their reputation and the reception that their products receive in the marketplace. TheySay can be used to analyse news reports, blogs, or postings on Facebook or Twitter - companies can get almost instant feedback.’
Karo Moilanen, CTO of TheySay, said: ‘The analytical depth and coverage of our technology leads to true Big Data-oriented sentiment analysis and opinion mining, and presents new ways to track and quantify deep affective signals in both unstructured social media and traditional structured text data.’
Tom Hockaday, Managing Director of Isis Innovation, said: ‘The Isis Software Incubator was set up to provide practical, commercial support and mentoring for ventures like TheySay, which has scope and potential across many industries.’
Alan Aubrey, Chief Executive Officer of IP Group, said: ‘TheySay’s contrarian approach to sentiment analysis is proving more accurate than traditional algorithms and could be of real interest to a broad range of companies looking to make strategic decisions based on, or aided by knowledge of, whether sentiment is generally positive or negative. We are excited about TheySay’s prospects and pleased to have made our first investment into a start-up from the Isis Software Incubator.’
TheySay received an Oxford University Challenge Seed Fund award to further develop the software in September 2010.