The project is in progress...

A tool for to develop mathematical models with chemometrics data

A set of algorithms for the purpose of performing sample predictions and classifications using spectroscopy data

Algorithms that can already be used:

Regression

• Partial least square regression (PLS)
• Support Vector Machine regression (SVR)
• Principal Component Regression (PCR)
• Random Forest.

Classification

• Principal component analysis + linear discrminant analysis (PCA-LDA)

Some details

When the algorithms in this tool have the search_hypermatameters method, that method can be used to search for the best set of hyperparameters based on the tool GridSearchCV of sklearn library.

In the run.py file, should be defined some variables like

• FOLDER_BASE = This is a string that define the directory where the output will be saved.
• ANALYSIS = This is a string that define the name of folder created for save output.
• save_results = If the tool should be save the results or not (is a bool variable)
• MAKE_AVERAGE = If necessary, samples are averaged. if True, it must be informed in which column the variables begin and the number of samples must be used to perform an average. (is a bool variable)