Feature‐level fusion of laser‐induced breakdown spectroscopy and Raman spectroscopy for improving support vector machine in clinical bacteria identification

Abstract

In clinical field, the diagnosis of many diseases and their development stages depend on the detection of the corresponding bacteria. Raman spectroscopy and laser‐induced breakdown spectroscopy (LIBS) are two novel spectral diagnostic technologies for clinical bacteria identification. Both of them have been used in clinical detection combined with optimized support vector machine (SVM). In this paper, two feature‐level fusion methods (before feature selection fusion [BFSF] and after feature selection fusion [AFSF]) were proposed to improve the performance of SVM classifier and reduce the analyzing time (including the parameter tuning, model training, and testing time) simultaneously by combining data of LIBS and Raman spectroscopy. Using the most important 10 feature lines as the inputs of the optimized classifier, the analyzing time could be reduced to 1 to 2 min and the correct classification rate (CCR) achieved 95.67%. Without optimizing SVM parameters, the two proposed methods could achieve rapid and accurate classification of pathogenic bacteria further. The AFSF method showed better results with less fusion features while the BFSF method achieved higher CCR at 100% with more features. Both two methods costed around 0.2 s for analysis. These indicate that the proposed feature‐level fusion methods can improve the performance of SVM for bacteria detection. Maintaining the highest diagnostic accuracy, they can achieve the minimum analyzing time.

Detection of low numbers of bacterial cells in pharmaceutical drug product using Raman Spectroscopy and PLS-DA multivariate analysis

“…In this work, we obtained similar values in terms of accuracy in comparison to previous reports showing sensitivities and specificities around 95-99% using RS coupled with multivariate analysis models. However, some approaches do not specify the limit of detection reached 45 or they presented theoretical detection limits below 100 CFU/ml based on RMSECV values obtained on the calibration and not in physical measurements 9 . Another common drawback is to ensure reproducibility of the results using a single model to detect samples with less than 10 CFU/ml or when cross-comparing among different species 22, 26, 35, 37 .…”

Detection of low numbers of bacterial cells in a pharmaceutical drug product using Raman spectroscopy and PLS-DA multivariate analysis

“…In this work, we performed RS coupled with multivariate analysis models and obtained sensitivity and specificity values between 98.5–99%, which compare to previously reported values. However, some of these approaches either do not specify the limit of detection, 45 or extrapolate to a detection limit below 100 CFU per ml based on RMSECV values, failing to provide a real measurement. 9 Another common pitfall is to assure reproducibility of the results based on a single model when analyzing samples with low concentration (<10 CFU per ml), or cross-comparing among different species.…”

Combining surface‐enhanced Raman spectroscopy and mid‐infrared spectroscopy in a data fusion model to forensic differentiate the electrostatic copying paper

“…How to cite this article: M. Zhu, Y. Chen, J. He, R. Yi, J Raman Spectrosc 2023, 54 (10), 1084. https://doi.org/10.1002/jrs.6590…”

“…Data fusion analysis is a popular trend to explore the potential value of different data sets and interpret them more accurately through visual models. [8][9][10] It can complement each other, integrate the strengths of each data set and make subtle differences easier to detect and evaluate by allowing different data sets to collide effectively. Spectral data fusion is one of the hot topics in forensic science.…”