Text Mining: Time to Consider Its Value

  Business Intelligence   Strategy  

Undoubtedly over the years, you've heard about data mining, neural networks and advanced statistical algorithms that can show which products are purchased together or identify potential cases of insurance fraud. Data mining, the statistical analysis of numbers to identify patterns, is still chugging along. More tools have entered the marketplace, but companies have struggled to show real value in the area of data mining.

In this column, I am going to introduce you to text mining, which has been a largely silent cousin of data mining over the last few years. Text mining, which focuses on the analysis and pattern recognition of text in documents and fields, has the potential to demonstrate business value and return on investment much more quickly than data mining. From a business sense, it is much easier to understand and can reap savings purely in terms of manpower reduction.

From a business standpoint, text mining can primarily add value in the classification and verification of records and documents. Let's start with a classification example: You're an insurance adjuster assessing damage to a car. You write down the following comments on a claim report: "CRACKED WINDSHIELD; DENT IN FRT BUMPER; LT ON FRT RT SD OF CAR FACING BRKEN." This report gets sent to the central office, and an analyst reads through the claim and classifies the claim into a bucket based on the damage…

Please log in or sign up below to read the rest of the article.