From the recent statistics of RBI and world bank, there are 80.245 million transactions occurring each second from India to worldwide, and 1.06 billion transactions occurring each second all over the world which are generating a robust database, analyzing and identifying each money trail as fraud or not is nearly impossible to make it possible we have some concept called self-organizing map which reduces data dimensionality, SOM is designed to convert complex data matrix into two dimensional, Each data point in the data set recognizes themselves by competing for representation. SOM mapping steps starts from initializing the weight vectors. From there a sample vector is selected randomly and the map of weight vectors is searched to find which weight best represents that sample. Each weight vector has neighboring weights that are close to it. The weight that is chosen is rewarded by being able to become more like that randomly selected sample vector. The neighbors of that weight are also rewarded by being able to become more like the chosen sample vector. This allows the map to grow and form different shapes. Most generally, they form square/rectangular/hexagonal/L shapes in 2D feature space. the weight with the highest vector shows up different compared to other data points. in this way we analyze fraud transactions.

Application of Self Organizing map:

1.Prediction of fraud transaction (used by bank )

2.Project prioritization and selection

3.Failure mode and effects analysis

 
 

Outline/Structure of the Talk

total time for talk including QnA = 20min

2min - the introduction of the speaker and deep learning

3min - difference, and need for SOM and other unsupervised learning algorithms

5min - an architecture of the basic neural network leading to the architecture of SOM

5min - explaining, working of SOM on large bank dataset

2min - live model learning and conclusion of SOM analyses

2min - questions and answers

1min- Buffer time

Learning Outcome

1)Now you can also implement SOM leaning on large bank transaction which can reduce fraud transaction,

2)More understanding of deep learning and neural network

3)Good analytic power

Target Audience

Researchers,Bankers,Data engineer,Data analyst,Data Scientists

Prerequisites for Attendees

Eager to learn something new and more going into unsupervised deep learning, a basic idea on "how deep learning works" would give a boost to talk

Slides

schedule Submitted 1 year ago

Public Feedback

    help