Knowing Human Behavior
Social Behavior Filtering (SBF)
We have invented an entirely new artificial intelligence algorithm to analyse individual behaviour which we call Social Behavior Filtering (SBF).
“Our fundamental belief is that Human Behavior is universal and that for any business problem it can be used as basis for customer/user behavior.
We consider everyone to be fundamentally the same and to be reacting as part of a community that they belong to. If we can quantify the community we will be able to predict the behavior of an individual.”
0 1
The SBF is a Graph Based Algorithm that leverages professional, personal, mobile usage and social data of customers, their individual loan history (if available) and the behavioral patterns of the community they belong, to determine the credit lending risk of an individual.
0 3
We believe that instead of evaluating an individual’s behavior in isolation we have to look at it as a part of a larger community, and then use it to measure his/her value for the business.
0 2
Individuals are extremely dynamic/agile and their changing behavior patterns creates a ripple effect in A.I. model. Social Behavior Filtering algorithm, learns these ripple effects, adapts from them and uses the knowledge to evaluate other customers.
End-Points
What Can SBF Do?
0 1
Evaluate current customers
-
Customer evaluation based on business specific historical data
-
Evaluation of new customers yet to inter
0 2
Trend forecasting
-
Capture untapped market
-
Ability to forecast trends in the market
-
Will enable focused marketing approach
0 3
Risk assessment
-
Detect identify fraud verify application data evaluate credit risk
-
Predict attrition based on the behavi
Knowing Human Behavior
Social Behavior Filtering (SBF)
0 1
Live System
​As soon as there is a individual behavior change, model improves.
0 2
Self Adapting
​As new data keeps coming in, the system itself will adapt.
0 3
Less influenced by variability in data
​Even if data is missing it can perform predictions.
0 4
Prediction about missing data
​In case of missing or unavailable data it offers prediction about that data.
0 5
Scalable
​Because the algorithm can be computed as part of parallel processes it is can be scaled with Big Data infrastructure to compute millions of customers.