Fraud-Detection

Fraud detection using Self Organising Maps

Unsupervised Learning.

The dataset file concerns credit card applications. All attribute names and values had been changed to meaningless symbols to protect confidentiality of the data. We need to find the outliers in the dataset i.e finding the customers who has cheated while the credit approval. First column corresponds to the customer ID, while the last column says whether the customer got the approval or not. We use Unsupervised Learning approach to detect the outliers.

Finding the outliers by plotting Mean Inter Neuron Distance.

Higher the MID, color will be lighter. Therefore they are the outliers.

After building and running the model, table shows the customers with ID who have attempted Fraud and their Fraud probabilities are as follows.