Project Apati
Developers:
Ibrahim Abdul Hammed, Shrenik Bhatt, Malhar Shah, Mujtaba Khan
Project Inspiration
From January 2014 to December 2017, Canadians lost more than $405 million to fraudsters
Fraud detection is a top priority for all indiviuals
Our Solution
We wanted to develop technology that detects and filters suspicious transactions based on user profiles
Transactions that are deemed consistent are validated automatically
Suspicious Activity Detection
High-Value Payments at cheaper location IDs are deemed suspicious
Suspicious Activity Detection
Transactions taking place in unusual locations outside of the norm are also flagged suspicious
scikit-learn Machine Learning Code
from sklearn import svm
import json
import pickle
import random
def train(cat_to_data):
category_to_model = {}
multiplier = 1
for cat in cat_to_data:
X = cat_to_data[cat]["X"]
Y = cat_to_data[cat]["Y"]
clf = svm.SVR()
clf.fit(X, Y)
category_to_model[cat] = clf
return category_to_modeldeck
By Malhar Shah
