The telecommunication company functioned in the market with extremely high competitiveness. Attracting new customers needs 5-10 times more expenses than maintaining an existing one. As a result, effective customer churn management and analysis of the reasons for customer churn are vital tasks for telecommunication operators. As a result, predicting subscriber churn by switching on the competitors becomes very important. Data Science and machine learning create enormous opportunities for solving this task to evaluate customer satisfaction with company services, determine factors that cause disappointment, and forecast which clients are at a greater risk of abandoning and changing services suppliers. A company that implements data analysis and modelling to develop customer churn prediction models has an opportunity to improve customer churn management and increase business results. The purposes of the research are the application of machine learning models for a telecommunications company, in particular, the construction of models for predicting the user churn rate and proving that Data Science models and machine learning are high-quality and effective tools for solving the tasks of forecasting the key marketing metrics of a telecommunications company. Based on the example of Telco, the article contains the results of the implementation of various models for classification, such as logistic regression, Random Forest, SVM, and XGBoost, using Python programming language. All models are characterised by high quality (the general accuracy is over 80%). So, the paper demonstrates the feasibility and possibility of implementing the model to classify customers in the future to anticipate subscriber churn (clients who may abandon the company’s services) and minimise consumer outflow based on this. The main factors influencing customer churn are established, which is basic information for further forecasting client outflow. Customer outflow prediction models implementation will help to reduce customer churn and maintain their loyalty. The research results can be useful for optimising marketing activity of managing the outflow of consumers of companies on the telecommunication market by developing effective decisions based on data and improving the mathematical methodology of forecasting the outflow of consumers. Therefore, the study’s main theoretical and practical achievements are to develop an efficient forecasting tool for enterprises to control outflow risks and to enrich the research on data analysis and Data Science methodology to identify essential factors that determine the propensity of customers to churn.

The work’s aim is to research a set of selected mathematical models and algorithms that examine the data of a single payment transaction to classify it as fraud or verified. Described models are implemented in the form of a computer code and algorithms, and therefore can be executed in real-time. The main objective is to apply different methods of machine learning to find the most accurate, in other words, the one in which the cross-validation score is maximal. Thus, the main problem to resolve is the creation of a model that could instantly detect and block a given fraudulent transaction in order to provide better security and user experience. At first, we determine the classification problem: which initial data we have, how we can interpreter it to find the solution. The next part is dedicated to presenting the methods for solving the classification problem. In particular, we describe such approaches as Logistic Regression, Support Vectors Method (SVM), K-Nearest neighbours, Decision Tree Classifier and Artificial Neural Networks; provide the notion of how these methods operate the data and yield the result. At the end, we apply these methods to the provided data using Python programming language and analyze the results.

Competition between marketing strategies of enterprises shifts to the use of artificial intelligence and begins to be considered in the context of competition between Data Science projects. Therefore, the issue of developing methodology and building a model in a particular area is relevant, which will make the project quite effective and ensure the achievement of goals for the company. The banking services market has a certain specificity of consumer behaviour, so forming marketing strategies is a somewhat complex process. Thus, banks face the task of maintaining the loyalty of their existing customers and attracting new ones. This article aims to build a marketing strategy to attract new customers in the banking sector using Data Science tools. The result of the study is the construction of two econometric models of the different bank's credit products: cash loans and credit cards, which determine the influence of various factors on this process and helps to distribute the advertising budget between different types of advertising. Using the built model, it was determined that advertising campaigns directly affect the increase in the number of new customers in the bank and the overall growth of brand knowledge about the banking institution in society. In addition, the determined weights of each influencing factor helped form an advertising budget, which increased customer inflows by 12%, with an average ROI of 3.18. Taking all into account, the model had shown its effectiveness in organising the bank's advertising campaign when decisions were made using Data Science technologies. The results obtained based on the models give a fairly clear understanding of the factors influencing the inflow of new customers in the bank, which will model the distribution of the budget for advertising campaigns in future periods and predict their effectiveness. Competition in the country's financial sector is forcing banking institutions to use data science in their marketing activities.

No abstract

This article presents a method for conducting a marketing research aiming to evaluate statistics of a “Willingness-to-Pay” random variable distribution. Similarly, this approach can be used for evaluating minimal price a customer is ready to sell a good for. Since a general survey tends to bring bias into WTP evaluation, we suggest reducing psychological pressure while asking a single question “Would you buy this product for X amount of money?”. It was empirically shown that this information is enough to conduct an experiment and evaluate the characteristics of a population distribution. The algorithm is easy to use, however needs an expert control for gaining higher accuracy. Using tools of simulation modeling we assessed the level of bias of an experimentally obtained distribution statistics compared to a real population statistics. The algorithm helps predicting individual demand and total income level depending on a product pricing level.