Cm confusionmatrix ytest ypr labelsnp.unique ytest normalizetrue disp ConfusionMatrixDisplay confusionmatrixcm displaylabels[cluster cluster ] dispot cmapplt.cm.Blues plt.show The error matrix of the model on the test set The Confusion matrix clearly shows that our model is very good at identifying cluster and slightly worse at identifying cluster . Perhaps it is relat to our outliers and perhaps we have too little data or there is too little data in the data information that is essential to the model. There may also be an anomaly that causes customers to behave differently within a sub product group for some reason. However there is no doubt that the quality of our model is so high that we can confidently proce to further analysis.
Review of model results thanks to the InterpretML library After all these operations we can finally look at the model and try to draw some conclusions about the characteristics of our assortment. setvisualizeprovider InlineProvider ebmglobal ebm.explainglobal show ebmglobal Summary of the significance of Taiwan WhatsApp Number List model features The basic view is a summary of the features that were most important for the model when pricting the class. It allows you to assess which features are worth looking at. The distribution of the significance of the features is also important in our case none of the features dominates the others. productcategory The most important feature from the models point of view is productcategory.
It clearly indicates which product categories are more desirable by customers and which are the opposite. Score is a relative value significant within the feature of the current model and should be treat as such. Negative values increase the probability of belonging to cluster positive values to cluster and values close to are neutral for the decision. Density is how much data of that type value is in the set keep this in mind especially if its very low. nunitofmeasure An interesting example for analysis is the nunitofmeasure feature which was rank in the top in terms of significance.