Application of Dimensionality Reduction and Machine Learning Methods for the Interpretation of Gas Sensor Array Readouts from Mold-Threatened Buildings

Paper is in the scope of moisture-related problems which are connected with mold threat buildings, sick building syndrome (SBS) as well application electronic nose for evaluation different envelopes and materials. The machine learning methods used to analyze multidimensional signals important components e-nose system. These derived from a gas sensor array, which, together instrumentation, constitute hardware this accuracy classification correctness buildings largely depend on appropriate selection data analysis used. This paper proposes method using Principal Component Analysis, metric scaling Kohonen self-organizing map, unsupervised methods, visualize reduce dimensionality data. For final observations identification datasets arrays analyzing air threatened by mold, other reference materials, supervised such hierarchical cluster analysis, MLP neural network random forest were