Open AccessCOMPREHENSIVE EXPLANATIONS FRAMEWORK FOR MACHINE LEARNING MODELS TRAINED ON TABULAR DATA
Author(s) -
Arman Zakaryan
Publication year - 2021
Publication title -
shph gitakan teghekagir
Language(s) - English
Resource type - Journals
ISSN - 2738-2559
DOI - 10.54151/27382559-2021.1a-29
Subject(s) - distrust , transparency (behavior) , computer science , data science , work (physics) , artificial intelligence , machine learning , knowledge management , engineering , computer security , psychology , mechanical engineering , psychotherapist
In the light of technological and scientific advances of the past decademachine learning models are becoming an inseparable part of many businesses.One of the shortcomings of ML models is the lack of transparency, which mayresult in a number of problems: hidden biases in the model, customer distrust,low adoption and usage etc. To increase the trust among the customers, theexplainable techniques should find their way to the end customers in adigestible format. In this work, we will explore some ML explainabilitymethods and provide a framework for presenting them in a comprehensivemanner.