Izza is my personal data science and machine learning toolbox.
There is mainly personal function for data visualisation:
function name | presentation |
---|---|
pca_visualisation | a pca visualtion (2 components) |
missingData | a tables showing missing data by features |
camembert_plot | a camembert plot |
kohohen maps | a kohohen maps with percentage of target class by neuron |
activation_frequencies | activation frequencies for a kohohen map |
There is methods to do model evaluation :
function name | presentation |
---|---|
fun_precision | precision at a determined percentage using predicted probabilities |
fun_recall | recall at a determined percentage using predicted probabilities |
f_macro_score | scorer allowing to find whether some clusters contain enough precision and recall for the target class |
precision_macro_score | scorer allowing to find whether some clusters contain enough precision for the target class |
recall_macro_score | scorer allowing to find whether some clusters contain enough recall for the target class |
viable_clusters | function allowing to know the interessing clusters found using the f_macro score. |