Open hechth opened 7 months ago
We need to train our own model -> first part would be to start with an scipy SVM using computed fingerprints from matchms and then train a model.
Details on the Scipy implementation of an SVM are available here: https://scikit-learn.org/stable/modules/svm.html#implementation-details
Accuracy for training set should be within 20 points
Related papers: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3163042/ https://www.sciencedirect.com/science/article/abs/pii/S0169743920301416 https://www.sciencedirect.com/science/article/pii/S0021967323004028 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8430916/
We need to train our own model -> first part would be to start with an scipy SVM using computed fingerprints from matchms and then train a model.
Details on the Scipy implementation of an SVM are available here: https://scikit-learn.org/stable/modules/svm.html#implementation-details
Accuracy for training set should be within 20 points
Related papers: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3163042/ https://www.sciencedirect.com/science/article/abs/pii/S0169743920301416 https://www.sciencedirect.com/science/article/pii/S0021967323004028 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8430916/