Training 1. Login to RAMANMETRIX® Online
Here we show where to find our RAMANMETRIX® online version, how to login and which instance related to your dataset size you should chose.
Training 2. Uploading your Data to RAMANMETRIX®
In this video we shortly explain the expertise level setting, how to import the training dataset of edible oils and which raw data infos are visualized in RAMANMETRIX®.
Training 3. Pretreating Data with Despiking and Calibration
In this video we show step by step how to optimally pretreat our example data set. For this, we discuss the despiking and calibration procedure and how the question marks at the symbols can help you to get more information about a certain step.
Training 4. Spectra Preprocessing by Baseline Correction and Normalization
This video includes the description of the data preprocessing procedures, which include the baseline correction and the normalization. We show how to just specific spectral regions for the normalization can be selected and how the preprocessed spectra are visualized.
Training 5. Quality Check of the Spectral Data
Herein, we mention a special feature of our RAMANMETRIX® software – the quality check with which it is possible to filter out corrupted data and to improve with that the accuracy of the model. A detailed description of all available quality filters will follow in another video.
Training 6. Construct your Model
Here we show how to construct an appropriate model for the edible oil training dataset at expertise level 5 and shortly discuss the results which are depicted in form of a confusion table and various graphs, including the mean spectra as well as the scatterplot, loading plot and variance plot. In addition, we show how parameters and model can be exported and how a training data report with all information on pretreatment and preprocessing parameters as well as the modelling parameters and results can be printed and saved.
Training 7. Applying the Created Model to a Test Dataset
This video shows how to verify the quality of your previously constructed model by a test dataset. We demonstrate how the test dataset can be uploaded, how to apply the constructed model to the test data and how the results are visualized.