Combined with information technologies, such as machine learning and big data, the predictive modeling skills can now roll out better forecasts since it can factor in as many variables as possible in analyzing data.
On the business front, such improvement in data analysis, or data mining, can mean saving huge costs by accurately gauging future risks.
Very few businesses are, however, effectively making use of solutions merged with those big data and machine learning technologies in their day-to-day businesses, and especially so in the financial sector that have to have top-tier prediction models to prevent financial risks.
“Many think big data -- one of the biggest buzz words in the IT sector -- is somewhere out there, not within their companies,” said Sophie Eom, the CEO of machine learning-based data mining solution provider Solidware, adding “Financial institutions, banks, insurance firms, and credit rating agencies are actually sitting on the treasure trove of data, but they are tapping less than 10 percent of it while the rest is unexplored.”
The current risk management solutions of financial firms look into just a dozen information categories when selling insurance or lending money, such as a type of job, age, and marital status, while there are thousands of different data sets available to make an accurate forecast for who will likely go bankrupt or be in arrears, according to the 29-year-old CEO, who founded the startup with her husband Olivier Duchenne in August 2014.
The company’s DAVinCI LABS boasts an easy user interface, allowing users to see the results of data analysis in line graph and histogram with just several mouse clicks.
The solution’s automatic clustering, which groups data that have similar traits automatically, is a feature that cannot be found in any other data analytics solution, Eom said.
For example, the solution automatically categorizes the people who were aboard the Titanic according to their gender, age, and travel class.
The DAVinCI LABS solution, which proved to be effective in saving costs in a pilot project with AXA Korea, the local subsidiary of French insurer AXA, will be integrated into its system to screen new customers.
In the pilot project, the solution showed that it could have saved 8 billion won (US$7.1 million) of costs for the insurance company by preemptively eliminating financial risks.
Depending on the results of the partnership with AXA Korea, the insurance firm’s French head office can employ the analytics solution.
“Among many programs tested at AXA’s global offices, only Solidware’s solution has rolled out positive results in saving costs,” said Eom, who used to work at AXA Korea before founding the startup.Man behind big data solution
Solidware made it to the top 10 in its maiden participation in a Startups 2016 competition, a sideline event of the ICT Spring Europe startup conference held in Luxembourg in May.
It also flexed its muscles in an online data analytics competition on Kaggle, a platform where researchers and companies contest with their data analytic capabilities.
Predicting the survival rate of people aboard the Titanic, mentioned above, is an actual task of the ongoing competition.
With just a single minute of test run to calculate the survival rate on the platform, DAVinCI LABS, without any optimization work, ranked in the top 3.5 percent among 5,200 participants, according to Solidware.
Eom’s husband Duchenne is the man behind the development of the powerful solution.
A prominent artificial intelligence researcher who won the best prize for his thesis on computer vision at the Computer Vision Pattern Recognition conference in 2009, Duchenne attracted seasoned experts in AI and machine learning to join the Korean startup.
Co-founder Duchenne used to work as a postdoctoral fellow for AI research at Carnegie Mellon University.
He also worked as a researcher at Aldebaran Robotics, a robot company owned by Softbank of Japan, and at Intel Korea before establishing Solidware.
Like many new startups, the co-founders started from the bottom up. In order to save office rent fees, they worked and had meals at their home together with their employees in the early stage until the company was acquired by Yello Financial Group, a fintech firm alliance, in March 2015.
“Duchenne has always wanted to apply his expertise in technology to the real world instead of only doing research at a lab,” said Eom, adding it was lucky to have him in Solidware since he was the one who attracted talented, seasoned AI and machine learning researchers into the team.
The 2-year-old startup was able to make money from its early business stage -- unusual for a fledgling startup -- and is expected to see its sales further increase with the planned global business expansion.
The Korean startup is planning to tap into Southeast Asia where the predictive modeling business is less competitive and advanced analysis solutions are in great need for the growing financial and banking sector in the regions. Other destinations include Japan. The company is in talks with a Japanese financial firm for partnership.
In Korea, KB Capital, Shinhan Bank, Welcome Savings Bank and AXA are currently deploying Solidware’s product in analyzing the data of their customers and Hyundai Card and SBI Savings Bank running a pilot project with Solidware’s solutions.
CEO Eom said that she believes problem-driven thinking have played an integral role in the firm’s smooth voyage in the startup world where most are scrambling to break even, or, in a bad case, just waiting to be acquired by bigger firms.
“Rather than trying to do something with technical skills you have, you have to find some pain points in society and try to solve them with your own capability, or sometimes with the help of others” said CEO Eom, who studied chemistry and business management at Seoul National University and got an MBA degree from HEC Paris.
She said the company will try to deploy its DAVinCI LABS solution in other sectors including gaming and medicine.
In the future, the solution could be used to help, for example, doctors to know probability of a patient getting a certain disease, or economists to make more accurate prediction on economic conditions for the coming years.
By Kim Young-won (firstname.lastname@example.org)