HP and Text Mining

What is the pratical application of text mining:
The practical application of text mining really is the combination of structured and unstructured data. Text mining applications pull data from sources such as word documents and emails in the form of texts, filters the data, and then translates it into a format the can be analyzed and recorded. Without text mining, written texts and other unstructured data would really become worthless data sources. According to studies done on BI and data mining, businesses and other BI clients are looking more and more to unstructured data as a primary data source. Probably the most pratcial application of text mining would have to be marketing.

How do you think text mining techniques could be used in other businesses:
There is almost a limitless amount of applications for text mining in other businesses. The most obvious use of text mining for other businesses would be to analyze written customer reviews and/or comments. Essentially, Text Mining can be used anywhere where there is a direct and free form line of communication between an entity and its actors. In the past, only a human was able to read, translate, record, and respond to these lines of communications. Text mining allows these processes to be completed without any human assistance. This means that new divisions and processes within a company could become automated such as responses to customer inquiries.

What were HP’s challenges in text mining? How were they overcome:
The two major challenges that HP experienced in developing its text mining was dimensionality and data dispersal. Most customer databases contain structured information that easily be extracted. The challenge and difficulty lies not so much in the extraction of the unstructured data but the combining and analyzing the unstructured and structured data together. These challenges were overcome by SAS Text Miner’s use of singular value decomposition along side of a prebuilt internal dictionary and synonym lists. Using the mathematical model of singular value decomposition, Text Miner is able to take highly variable and dimensional data of written texts and reduces it to a lower dimensional form with noise and redundancy reduction. The use of the internal dictionaries further reduces and simplifies the data to a form where it can analyze it against structured data.

In what other areas, in your opinion, can HP use text mining:
HP can use text mining in other areas such as monitoring comments written on social media pages and blogs. HP can use text mining to extract data about itself from publication histories such as newspapers, magazines, and journals.