Each participant needs to provide idea with process flow on how to develop a predictive analytics model that will most accurately foresee and predict corporate customer loan default using the publicly available data (e.g. news, blogs, forums, websites, court/arbitration center judgements, social media, stock exchanges, SEBI, rating agencies etc).
Each shortlisted team will be given with 20 names of SBI Corporate Customers (chosen randomly) in the Hackathon Phase and the team will have to give the probability of default over next 6 months of these customers using publicly available data. Participants whose predictions come closest to the truest i.e. best correlated with real timeframe of 6 months will be awarded. The analytical models can then be used in the bak for generating Early Warning Signals. Hence, the winners will be declared 6 months after the submission of Hackathon Phase Model/Prototype.
Individual Teams are required to submit the Requirement and Technical Specifications of the product to be developed in Phase 1 (Ideation Stage). The ideas will be evaluated on the basis of the following:
The Hackathon will be conducted in 2 phases spanning over 5 weeks (tentatively).
|PHASE I||Ideathon Phase – Idea Submission||24.01.2019 – 07.02.2019|
|PHASE II||Hackathon Phase - Prototype Development||26.02.2019 – 26.03.2019|
Decision taken by SBI is final and binding. SBI may not award any prize if it does not find any satisfactory model.
|Name of the Judge||Status in the committee|
|Mr. Kamal Deep Prasad||Chairman|
|Prof. G Siva Kumar||Member|
|Mr. Pankaj Kumar||Member|
|Mr. Arvind Kumar||Member|
You can either participate as an individual or as a team.
Every team can have a maximum of 4 people
If you have any queries or feedback, please email us at firstname.lastname@example.org