Sberbank Gender Prediction

Task description

One of the most valuable sources of customer information is bank transaction data. In this competition, participants are asked to answer the question: is it possible to predict the gender of a client using information about outcome transactions and income on a bank card? And if so, what is the accuracy of such a prediction?

File descriptions

transactions.csv – transactional data on banking operations.
train set.csv – training set with client gender marking (0/1 - client gender).
codes.csv – the table contains description of MCC transaction codes.
types.csv – table contains description of transactions.
test set.csv – data set with 2656 client id that is independent of the training data set, but that follows the same probability distribution.

Data fields

transactions.csv

client_id – client is id;
datetime – transaction date (format - ordered day number hh:mm:ss - 421 06:33:15);
code – MCC transaction code;
type – transaction type;
sum – sum of transactions: "+" — accrual of funds to the client (incoming transaction), "-" — write-off of funds (outcoming transaction);

train_set.csv

client_id – client is id;
target – client gender;

codes.csv

code – MCC transaction code;
code_description – MCC transaction code description;

types.csv

type – transaction type;
type_description – transaction type description;

test_set.csv

client_id – client is id;