Kuljus, Kristi, juhendajaKoloydenko, Alexey, juhendajaViigand, SiimTartu Ülikool. Loodus- ja täppisteaduste valdkondTartu Ülikool. Matemaatika ja statistika instituut2024-06-282024-06-282024https://hdl.handle.net/10062/100425The aim of the master’s thesis is to study and set reference results for assessment of residual tumor margins of basal cell carcinoma. The analysis is based on Raman microspectroscopy measurements of tissue samples extracted during Mohs surgery. Logistic regression, linear discriminant analysis and quadratic discriminant analysis are used to develop classification rules, several set-ups of spectral feature variables are considered. The best classification result is achieved with the logistic regression model with 30 original spectral features. The best model is validated on a test set of extracted tissue samples. Concerns regarding the data can be taken into account and the obtained reference results can be used in future analyses, when models of higher complexity could be studied for classification.enAttribution-NonCommercial-NoDerivs 3.0 Estoniabasal cell carcinomaRaman spectroscopyclassificationlogistic regressiondiscriminant analysisbasalioomRamani spektroskoopiaklassifitseeriminelogistiline regressioondiskriminantanalüüsmagistritöödvõrguväljaandedThesis