For the confusion matrix to be created there needed to be a comparison between the predicted categories, i.e., the categories given by the AI Classifier and the actual categories, i.e., the manual categorization that was done by us. So, the TP, FP, TN and FN values were calculated for each of the patent numbers using the actual and the predicted values.

We considered 4 performance metrics scores:

• Accuracy: The ratio of correctly classified points (prediction) to the total number of predictions.
• Recall: The ability of a model to find all the relevant cases within a dataset.
• Precision: The ability of a classification model to identify only the relevant data points.
• F1 score: The harmonic mean of precision and recall. It takes both false positive and false negatives into account. Therefore, it performs well on an imbalanced dataset.