I have a strong background in mathematics and data analysis. My expertise lies in handling large and complex data structures using advanced methods based on latent variables. I have successfully implemented these techniques in various industry and research projects.
Education
2005 - 2009
Experience
2010 - 2015
Skills
Proficient in Python, R, and MATLAB
Latent Structure Analysis
1 Latent Math. An Introduction
Latent variables/score vectors. Examples. Linear projections. OLS solution. Ridge Regression. PLS Regression. Multi-step forecasting. Model validation procedures.
2 Industrial solutions
Analysis, control, and supervision of processes in complex data paths.
3 Applied research projects
Multiple, multi-step forecasting with optimal classifications
4 Data handling
Large and complex data structures are arranged so that multi-step forecasting can be carried out. It gives new methods in Path Modelling.
5 Data analysis
Data are analyzed by forward and backward procedures. Optimal solutions to Industry Standards requirements are obtained.
6 Efficient methods
Principles of finding optimal balance between fit and precision are used to obtain stable numerical results for generalized inverse, QR decomposition and others. For multi-way data directional inverses are defined, which gives natural extension of methods for two-way data (matrices) to multi-way data.
Most commonly used regression methods are carried out by the same algorithm. It means that popular methods in one regression can be applied to others. For instance, graphic analysis of variation in data in PLS Regression can be applied to other regression methods.
7 Advanced techniques
8 Common errors made in analyzing low rank data
10 common errors are presented, where the user of the software presents invalid or wrong conclusions from the analysis. This is especially important in the applications of Path Modelling, where serious errors in conclusions are common.
