-  SK Planet T Store to leverage user review analyses for optimal content recommendation 
-  Utilizes opinion mining to quantitatively analyze expressions of sentiments to extrapolate relevant
    content attributes 
-  Service available from Feb. 6 in T Store’s game category and to expand across all types of T Store
    contents over time

Seoul, Korea, February 6, 2014—SK Planet today announced the launch of a content recommendation service using
big data, the first of its kind in Korea.
 
The newly released service applies opinion mining, a big data analysis technology, to T Store’s customer reviews to
identify relevant content attributes. These are then used as a basis for endorsing contents to potential buyers.
 
Opinion mining is a big data processing technology involving the collection of information available online on a topic
of interest and then analyzing to extrapolate relevant sentiments and views. The advantage of opinion mining is that
it breaks down web documents into the smallest units for analysis, which covers not only hard facts but also subjective
elements, such as sentiments expressed in words.
 
SK Planet’s opinion mining technology has evolved since the R&D began in 2008. To raise analytical accuracy,
documents are categorized by keywords, attributes, and positive/negative sentiments, etc. followed by linguistic
processing and analysis, which became the basis of a ‘dictionary of emotions’ for quantifying the degree of
expressed emotions. This proprietary dictionary identifies comprehensive attributes for a given topic, providing
for a more accurate analysis.
 
The big data driving the analysis is the customer reviews on T Store. In addition to the information offered by
the content seller, the buyer is able to base her decision on reviews put up by others who have already experienced
the content. The benefit is the enhanced quality and quantity of information to assist in selection.
 
For the consumer this is great. His choice will be supported by user reviews in addition to download ranking, price
and release date. Purchase decisions will be smarter and better informed as a result.
 
Recommendations based on opinion mining data will initially be offered in T Store’s game category. The level of
satisfaction of game users as indicated in their reviews will be quantified on a scale of five criteria: enjoyment,
graphics and sound, technical quality, control and game elements. Games with high scores will be featured
prominently to assist consumer choice. Recommendations will be updated weekly. 
 
Lee Jae-hwan, Head of Digital Content Business Division of SK Planet explained, "We have developed the
content recommendation service based on opinion mining in order to help our users quickly and easily find
their pick in a sea of mobile contents. T Store is always committed to bringing variety, fun and innovation
to our users.”