BI (business intelligence) vendor SAS Institute Inc. has signalled a move toward analysing unstructured data after it announced last week a deal to use Inxight Software Inc.’s text-mining technology.
The technology from Santa Clara, Calif.-based Inxight, a wholly owned subsidiary of Xerox Corp. that focuses on text mining, will be available in SAS Text Miner, due midyear.
Specifically, SAS plans to use Inxight’s LinguistX Platform, a natural-language text solution for analysing words, phrases, and sentences, and Inxight’s Thing Finder, which identifies and extracts from documents key content such as company names, products, people, addresses, and dates.
Text Miner will take advantage of Inxight’s technology for analysing documents that contain text in a variety of languages.
Furthermore, Inxight’s capabilities will enable Text Miner to access and run analysis against as many as 15 types of data files, including unstructured data sets such as e-mail, Microsoft Office applications, and PDF and HTML files.
“We think that text mining has broad implications across several industries,” said Wayne Thompson, product manager for Enterprise Miner at SAS in Cary, N.C.
Thompson said that Text Miner is part of a broader strategy at SAS to analyse both structured and unstructured data. “We provide the analytics to do the clustering, predictive modelling, and CRM activities,” he said.
Examples of text-mining use in the enterprise include HR applications that filter and match resumes to specific job openings and portals that are capable of pushing out specific text documents to users based on profiling, Thompson said.
Philip Russom, an independent industry analyst based in Waltham, Mass., said that it is a natural fit for BI vendors to partner with search and text-mining vendors.
“The BI players’ real expertise is in structured data, not unstructured data,” Russom said.
For BI purposes, text mining is much more efficient than search technology, according to Russom.
“A dumb old search engine just goes through and looks for words, but it doesn’t know what they mean,” Russom said.
As a result, all that most search engines do is collect and alphabetize documents based on a specific word, phrase, or string, Russom added. Text miners, on the other hand, can search for that specific string as well as related words.
“A text-mining tool is lot smarter, it can make associations based on a library of synonyms,” Russom said.
Smarter Search Engines
Inxight’s text-mining technology has powerful advantages when compared to BI search engines.