Pre-trained Predictors

Smart Bot provides model to predict probable outcomes with a confidence score, for better informed decision making. Some of the most common pre-trained predictors are available with Smart Bots. 

• Loan Credit Fraud Propensity: This will predict the loan credit default propensity of the loan application based on various parameters like Application number, Gender, Marital Status, Number of Dependents, Education, Date of Birth, Employment Type, Industry, Applicant Income, Loan Amount, Loan Term, Credit History, Property Area, Loan Type, Interest Rate Type and KYC Required. The home loan prediction dataset was modified to create a dataset to trian this predictor, which can be located at: https://www.kaggle.com/gavincanacam/home-loan-predictions#Train_Loan_Home.csv
• Credit Card Fraud Detection: This will predict the probability of the fraudulent credit card transactions based on amount, time and other principle components. It is important that credit card companies are able to recognize fraudulent credit card transactions so that customers are not charged for items that they did not purchase. The datset used to train this predictor can be found at: https://www.kaggle.com/mlg-ulb/creditcardfraud.
• Insurance Claim Prediction: This predictor will predict the probability of health insurance being claimed or not based on policy holder’s age, gender, BMI (body mass index), number of children/dependents, smoking state, region, charges. This is "Sample Insurance Claim Prediction Dataset" which based on "Medical Cost Personal Datasets" to update sample value on top. The dataset used to train this predictor can be located at: https://www.kaggle.com/easonlai/sample-insurance-claim-prediction-dataset.
• Mobile money transaction fraud detection: This will predict the probability of the fraudulent mobile money transactions based on various parameters like transaction time (mapped into steps here), type of transaction (cash-in, cash-out, debit, payment and transfer), the id of customer who started the transaction, initial and new balance of customer, id of recipient, initial and new balance of recipient. There is a lack of public available datasets on financial services and specially in the emerging mobile money transactions domain. Part of the problem is the intrinsically private nature of financial transactions, that leads to no publicly available datasets. The dataset used to train this predictor can be located at: https://www.kaggle.com/ntnu-testimon/paysim1.