Communication of machine learning results through tabular outputs

The field of data mining constantly pursues the development of predictive models, crucial for various applications. While much emphasis is placed on improving the accuracy of these models, their comprehensibility to analysts and end-users is an often overlooked aspect. This thesis addresses this gap by presenting an empirical study that investigates the interpretability of alternative representation format, with a specific focus on decision trees and decision tables. In our research, decision trees are identified as challenging to read and understand, prompting the exploration of a solution. Motivated by the objective of enhancing model interpretability, we propose the development of a Python program designed to convert complex decision trees into more accessible decision tables. Decision tables, known for their clarity and ease of interpretation, emerge as a favorable alternative.