Rapid and Low-Cost Detection of Thyroid Dysfunction Using Raman Spectroscopy and an Improved Support Vector Machine

This study presents a rapid and low-cost method to detect thyroid dysfunction using serum Raman spectroscopy combined with support vector machine (SVM). The serum samples taken from 34 thyroid dysfunction patients and 40 healthy volunteers were measured in this study. Tentative assignments of the Raman bands in the measured serum spectra suggested specific biomolecular changes between the groups. Principal component analysis (PCA) was used for feature extraction and reduced the dimension of high-dimension spectral data; then, SVM was employed to establish an effective discriminant model. To improve the efficiency and accuracy of the SVM discriminant model, we proposed artificial fish coupled with uniform design (AFUD) algorithm to optimize the SVM parameters. The average accuracy of 30 discriminant results reached 82.74%, and the average optimization time was 0.45 s. The results demonstrate that the serum Raman spectroscopy technique combined with the AFUD-SVM discriminant model has great potential for the detection of thyroid dysfunction. This technique could be used to develop a portable, rapid, and low-cost device for detecting thyroid function to meet the needs of individuals and communities.