Treffer: Machine learning for compositional data analysis in Support of the Decision Making Process

In recent years, the development of digital technologies brings alot of changes in the way of operating, leading, and working processesin companies. Accordingly, advanced technologies such as ArtificialIntelligent, Big Data, Internet of things, etc., are widely applied toaggregate, transform, and analyze data, thereby inferring meaningfulinformation from the results, making important decisions. As a branchof AI, machine learning (ML) is a method of data analysis that constitutesanalytical model-building automation. The main objectives ofML are designing algorithms that can learn from data by themselves,identify patterns, and adapt them without human intervention. Thegoal of this chapter is to summarize the researches related to applyingML to compositional data (CoDa), including principal componentanalysis (PCA), clustering, classification, and regression. CoDa is aspecial type of data, well-defined on the Simplex space. Since it carriesonly relative information, the traditional methods can not be applieddirectly to this type of data without adapting or transforming datainto normal form. Besides, we will introduce a transformation methodbased on Dirichlet density estimation to transform CoDa into real dataand apply those transformed data in anomaly detection using SupportVector Data Description (SVDD). A simulation example to illustratethis method is also provided at the end of the chapter.