Correlation–Comparison Analysis as a New Way of Data-Mining: Application to Neural Data

• Published in: SN computer science 2023-09, Vol.4 (5), p.636, Article 636
• Main Authors: Grbatinić, Ivan, Krstonošić, Bojana, Srebro, Dragana, Purić, Nemanja, Dubak, Marija, Dušanić, Vladan, Kostić, Vladimir, Milošević, Nebojša
• Format: Article
• Language: English
• Subjects: Classification; Computer Imaging; Computer Science; Computer Systems Organization and Communication Networks; Correlation; Correlation analysis; Data mining; Data Structures and Information Theory; Information Systems and Communication Service; Multidimensional data; Original Research; Parameters; Pattern Recognition and Graphics; Samples; Software Engineering/Programming and Operating Systems; Vision
• ISSN: 2661-8907; 2662-995X; 2661-8907
• DOI: 10.1007/s42979-023-02086-4

This paper aims to present a way of multidimensional data-mining termed correlation–comparison analysis (CCA). It was applied to neural data to demonstrate its utility in neuron-classification problem. The CCA represents a semi-quantitative way of inter-sample comparisons. The methodology comprises the generation of inter-parametric correlation and alpha-error ( p value) matrices. The main step is p -comparison for the same parametric pair defined between the two samples. This comparison has a semi-quantitative binary character that does not involve issues, such as false discovery rate (FDR) in multiple comparisons. As a result, the outcomes obtained are: (1) a correlation match, (2) a correlation mismatch of the first kind, the main type of a correlation mismatch, (3) a correlation mismatch of the second kind, the strongest one but very rarely observed in biological systems and obtained on a very small number of parameters. The correlation mismatch of the first kind is the target mismatch, i.e., the mismatch of tracing interest and represents the very reason why the study itself is performed. The application of the CCA led to the effective neuromorphofunctional classification of caudate interneurons into appropriate clusters and their feature-based description. The CCA analysis is a multidimensional bi-sampled classification tool that can be very useful for similar samples to explain their differences. Graphical Abstract Correlation–comparison analysis is based on the following transcorrelative relations between correlations belonging two similar samples: (1) the particular inter-parameter correlations are matched between the two samples in terms of their statistical insignificance in both of the samples, or in terms of accomplished statistical significance in both of the samples and both of the correlations are characterized with the same direction, (2) correlations are mismatched in terms of accomplished statistical significance in one of the samples but not in the other one (correlation mismatch of the first kind, the main type of correlation mismatch), (3) correlations are mismatched in terms of accomplished statistical significance in both of the samples but the correlations are characterized with the opposite direction (correlation mismatch of the second kind, the strongest one but very rarely observed in biological systems and obtained for a very few number of parameters).