Samasati Cocina Pvt. Lmt.

Now the internet is a booming industry and wherever there is good, there is also bad, being that said we know that internet brings prosperity to various kinds of businesses and small scale industries through the medium of E-Commerce, it also started this whole new thing of online transaction frauds. Now the concept we are dealing with right now is that sometimes there are users trying to temper the payment getaways used in various E Commerce sites. So what we are doing right here is developing an API which can be used at the backend of the website of your E Commerce organization and our API will keep an eye on each and every transaction request made on your payment gateway and before processing any request on your end, our API will analyze all the details of the request and based upon the previous data, it will predict weather the request is genuine or someone is trying to temper or hack the site. Approach to deal with Problem:

• Do exploratory data analysis on each and every column and understand the impact of each feature.
• Build a baseline model.
• Create new features and see whether those features improve the performance of the model. If yes, retain the features or else drop those features. I used forward feature selection for that.
• Iteratively continue this approach until we get a good score.