•    prediction model – based on neural networks (Deep Neural Networks)

◦   the use Python language tools, including Pandas, Numpy, Tensorflow and others.

◦   the use of statistical analysis to classify the client population

◦   take the history of each client into account

◦   make it possible to predict the services they might be interested in

◦   allows you to acquire a set of clients who may be interested in a specific service

•    exploratory analysis – emphasis on visually and numerically tracked regularities in client behaviour, takes into account differences based on statistical indicators.

•    use of logical arguments – a high level of interpretation

•    strongly intuitive prediction model parameters

•    meaningful testing of the sensitivity of a model to parameters

•    classification methods – for thorough testing and comparison of models