COMPANY TECHNOLOGY
Our R&D History
Up to 2007 (before establishing the company):
We developed a system for collecting various network data on the Web, such as from the blogsphere, social network sites, and Internet auction sites, and then analyzed the large-scale complex network structure of this data. We then constructed a mathematical model to explain the underlying mechanism responsible for the universal characteristics associated with the power law distribution. Furthermore, we researched and developed an algorithm for extracting user and item groups with similar preference characteristics, in the form of a virtual community. This process gave rise to a discovery engine which provides recommendations for the information that the user desires regardless of the data form.
From 2007 onward (after establishing the company):
We implemented the discovery engine as a system solution aimed at corporate services. We constructed an information system platform with the capacity to simultaneously offer the service to hundreds of companies. We also moved the system to a larger datacenter.
We developed an architecture that uses user click information generated from within the actual system as feedback to improve the accuracy of the calculation. This is a scalable system in which increased use of the system results in increased self-learning.
From 2009 to the present:
In preparation for being able to handle a larger scale dataset in the future, we developed a system that can be scaled out. We constructed a platform that can reliably handle one billion records of behavior history data traffic per day. In addition, we expanded the database system through virtualization, making it possible to process enormous amounts of data and serve as a universal Behavior Warehouse.
Governmental R&D grants received in the past:
1. “Original Technology-Seeds Development Program”, Japan Science and Technology Agency.
2. “Grant-in-Aid for Young Scientists”, Japan Society for the Promotion of Science.
3. “Grant-in-Aid for Exploratory Research”, Japan Society for the Promotion of Science.
4. “Consortium R&D Project for Regional Revitalization”, New Energy and Industrial Technology Development
Organization of Japan.
5. “Grant-in-Aid for JSPS Fellows”, Japan Society for the Promotion of Science.
PATENT
Scigineer has filed six patents for the following concepts. Deqwas can provide our recommendations on various formats. Major formats are:
・ Recommends adequate items by detecting user communities whose preference is similar to each other based on
purchase history.
・ Detects trend leaders/followers from fashion propagation dynamics.
・ Recommends information based on dynamically changing popularity.
・ Navigates users into their mileage partners based on their loyalty.
・ Finds out better items and navigates users to those EC sites.
・ Transforms user behaviors into graphs and detect each community that has similar tastes and recommends
items/users based on their taste
AWARD
Red Herring
|
|
||||
Red Herring
|
|
||||
CNET Japan, TechVenture2009
|
 |
||||
|
2nd prize at Infinity Ventures Summit, 2009 Spring Launch Pad
http://www.infinityventures.com/ivs/archives/