뉴스
KUBS 소식
['Brightics AI' Information Session] A Fast and Easy Big Data Analysis Platform
2018.04.04 Views 2250 경영대학
['Brightics AI' Information Session]
'A Fast and Easy Big Data Analysis Platform'
'A Fast and Easy Big Data Analysis Platform'
On April 3 (Tue) 5pm, KUBS had an information session for a big data analysis platform, ‘Brightics AI’ at B202, Hyundai Motor Hall. Brightics AI is a big data analysis platform developed by Samsung SDS. It is equipped with a prescriptive algorithm that provides solutions based on its AI analysis models across industries including manufacturing, marketing, and logistics.
Earlier on the information session, Professor James Jungho Pak from College of Engineering introduced Brightics AI. “I am sure this big data analysis platform by Samsung SDS will be a useful tool for researches in universities”, Professor Pak said. “We had the information session last time at the Science and Engineering campus during winter vacation. This time here on Humanities campus.”
The information session was presented by Ms. Eun-Joo Lee, head of Samsung SDS Data Analysis LAB. “Normally it took a very long time processing data at offices due to its terabyte sized data. However, the development of analysis platform makes it possible to fast process big data, only by the machine work.” Ms. Lee said. In consequence, corporations now prefer this route of data analysis. Python is the most preferred program rather than SAS (Single Attachment Station). In addition, whereas offices had only one person to analyze the data, nowadays they have a team set for the analysis, composed of IT leader, data analyst, C level, data scientist. Ms. Lee added, “You can import functions from SAS or Python. So Brightics is basically an integration platform that delivers analyzed data and shares techniques for analyzing”.
Ms. Eun-Joo Lee then shared an example of analysis by Brightics. The task was to find a player who is capable of playing the closest to Christiano Ronaldo, using data of 17,000 professional soccer players. With numerical data of 30 variables including potential, power and overall score, the data were categorized into 7 groups by similarity. Through further analysis on smaller group, they were able to narrow down the list to a player whose play is the closest to Ronaldo. “In fact, these are the techniques used in real life. The result of the analysis can be presented to sports team owners.” Ms. Lee added.
Brightics can be used for data analysis in a range of industries for predicting△signs of abnormality in production △optimal duration of component △delays in shipping △sales prediction △market reaction △monitor feedbacks △number of customers △diagnosis of arrhythmia using S-Patch △management of electricity usage. Ms. Lee concluded, “Brightics is equipped with a prescriptive algorithm that provides solutions based on its analysis, which non-professional analysts have an easy access to. It will be distributed to Korea University for users’ free access”.