Investigation of Trees Algorithms for Improving Quality of Client's Data in Data Mining Tasks

摘要

Machine learning methods, which are tree-based algorithms, have found their application in the task of improving the quality of client data. Among such methods are distinguished: decision trees and decision forests, random trees, ensembles of trees and forests. Such methods are used both to save large amounts of data from redundant information, to restore missing values and to determine anomalous values. Subsequently, on the data transformed by such algorithms, it is possible to build highly accurate forecasts of changes in the target parameter, whether it is the size of profit or investment. Thus, the effectiveness of applying algorithms on trees in a new field for the introduction of machine learning is proved.