学术文献库

Although Raman spectroscopy is widely used for the investigation of biomedical samples and has a high potential for use in clinical applications, it is not common in clinical routines. One of the factors that obstruct the integration of Raman spectroscopic tools into clinical routines is the complexity of the data processing workflow. Software tools that simplify spectroscopic data handling may facilitate such integration by familiarizing clinical experts with the advantages of Raman spectroscopy. Here, RAMANMETRIX is introduced as a user-friendly software with an intuitive web-based graphical user interface (GUI) that incorporates a complete workflow for chemometric analysis of Raman spectra, from raw data pretreatment to a robust validation of machine learning models. The software can be used both for model training and for the application of the pretrained models onto new data sets. Users have full control of the parameters during model training, but the testing data flow is frozen and does not require additional user input. RAMANMETRIX is available in two versions: as standalone software and web application. Due to the modern software architecture, the computational backend part can be executed separately from the GUI and accessed through an application programming interface (API) for applying a preconstructed model to the measured data. This opens up possibilities for using the software as a data processing backend for the measurement devices in real-time. The models preconstructed by more experienced users can be exported and reused for easy one-click data preprocessing and prediction, which requires minimal interaction between the user and the software. The results of such prediction and graphical outputs of the different data processing steps can be exported and saved.