Prediktivní model odchodu zaměstnanců

This thesis deals with the practical use of statistical methods in data mining in the area of prediction of employee turnover. The first part of the thesis describes the company's starting situation and the necessary theoretical basis from the area of employee turnover and retention. The causes and consequences of the fluctuation are described here with the indicated measures used to increase employee retention. This section uses data from personnel systems and questionnaire surveys in the context of exit interviews. In the second part of the thesis, theoretical principles are processed when managing a project in data mining. At the same time, various statistical methods are described with explanations and possibilities for evaluating their deployment. In the practical part, the above theoretical concepts are applied to the project predicting the departure of employees as binary classification tasks. After conducting an exploratory analysis of the data, it is possible to select a suitable model based on the results from the statistical SW. The selected model has been validated on last year's data. The last section described the results of the model and the appropriate recommendations for further use in the practical life of the company.