A New Type of Classification Algorithm Inspired by the Chromatographic Separation Mechanism

Currently, one of the important problems related to data mining is the processing of large data sets. The article presents an algorithm that can be used in issues related to the classification of large-volume data sets. The motivation to define this type of algorithm was the fact that currently the methods used to process this type of data are subject to several significant limitations. The first significant limitation in the use of classical classification methods is the need to ensure a constant data size. The second type of limitation is related to the dimension of the data. The last type of limitation that occurs when using classic classification algorithms is related to the situation that a given input vector may contain data belonging to many classes at the same time, then we are talking about the so-called multiclass vectors.The work attempts to define a data classification algorithm inspired by the method of chromatographic separation of chemical substances. This method is widely and successfully used in analytical chemistry. The article presents the results of calculations for sample data sets and discusses issues related to the properties of the defined algorithm, which concern the algorithm training process and the classification of single-class, multi-class and variable-length data vectors.