Australian text analytics company Leximancer claims a huge leap forward in the ability for its software to understand what customers are saying.

Leximancer version 3.5 contains a new Automatic Sentiment Lens feature which will identify whether customers are reacting positively or negatively, and eliminate much of the ambiguity in results returned by other ‘Voice of the Customer’ solutions.

"Sentiment is generally modelled via the introduction of a list of typical sentiment, or emotive, seed words to help 'steer' the technology to determine, or profile, sentiment in the text," said Leximancer Chief Scientist, Dr Andrew Smith.

"However, this can be prone to error due to the ambiguity of human language and context. For example: is the term 'rubbish' in the text referring to waste material or poor product quality?" Dr Smith said.

"We have extended our learning algorithms to positive and negative terms, to better define their usage and context - and dramatically reduced the impact of rogue or ambiguous sentiment terms in the process."

The Brisbane-based company said Leximancer software is being increasingly used by organisations in Australia and overseas to discover meaning in the vast quantities of text that defy traditional approaches to data mining.

Leximancer partner SEMA helps its customers improve customer and business processes by leveraging the resulting critical data in multichannel communications. SEMA’s Dave Kerr says that Leximancer has been a genuine support to its entry into the field of unstructured data analytics.

"Leximancer is simple and intuitive to use and we’re able to produce insights within a matter of minutes. It also provides the flexibility for us to really delve into the data to extract meaning," Mr Kerr said.

"We’ve been caught in the past with tools that don’t offer as much flexibility as we require which compromises our ability to service the market. We don’t have that problem with Leximancer."

Leximancer General Manager, Mac Exon-Taylor, said the response from SEMA was typical of what customers were saying about Leximancer software.

"Customers are becoming increasingly knowledgeable and understand that many of the claims made by text analytics companies need to be tested," Mr Exon-Taylor said.

"We find that organisations which explore what our software can tell them about their data - without large set-up costs and time - are amazed.

With the Automatic Sentiment Lens they can now utilise a whole host of data held within and outside the organisation, including social media, to get a much more complete picture of what people are saying about them," Mr Exon-Taylor said.

Leximancer Pty Ltd was established in 2006 by The University of Queensland's main research commercialisation company, UniQuest Pty Limited.